Obesity and type 2 and diabetes mellitus (T2D) are two dual epidemics whose shared genetic pathological mechanisms are still far from being fully understood. Therefore, this study is aimed at discovering key genes, molecular mechanisms, and new drug targets for obesity and T2D by analyzing the genome wide gene expression data with different computational biology approaches. In this study, the RNA-sequencing data of isolated primary human adipocytes from individuals who are lean, obese, and T2D was analyzed by an integrated framework consisting of gene expression, protein interaction network (PIN), tissue specificity, and druggability approaches. Our findings show a total of 1932 unique differentially expressed genes (DEGs) across the diabetes versus obese group comparison (p≤0.05). The PIN analysis of these 1932 DEGs identified 190 high centrality network (HCN) genes, which were annotated against 3367 GO terms and functional pathways, like response to insulin signaling, phosphorylation, lipid metabolism, glucose metabolism, etc. (p≤0.05). By applying additional PIN and topological parameters to 190 HCN genes, we further mapped 25 high confidence genes, functionally connected with diabetes and obesity traits. Interestingly, ERBB2, FN1, FYN, HSPA1A, HBA1, and ITGB1 genes were found to be tractable by small chemicals, antibodies, and/or enzyme molecules. In conclusion, our study highlights the potential of computational biology methods in correlating expression data to topological parameters, functional relationships, and druggability characteristics of the candidate genes involved in complex metabolic disorders with a common etiological basis.
Citation: Abdulhadi Ibrahim H. Bima, Ayman Zaky Elsamanoudy, Walaa F Albaqami, Zeenath Khan, Snijesh Valiya Parambath, Nuha Al-Rayes, Prabhakar Rao Kaipa, Ramu Elango, Babajan Banaganapalli, Noor A. Shaik. Integrative system biology and mathematical modeling of genetic networks identifies shared biomarkers for obesity and diabetes[J]. Mathematical Biosciences and Engineering, 2022, 19(3): 2310-2329. doi: 10.3934/mbe.2022107
Related Papers:
[1]
Sandro Claudio Lera, Didier Sornette .
Evidence of A Bimodal US GDP Growth Rate Distribution: A Wavelet Approach. Quantitative Finance and Economics, 2017, 1(1): 26-43.
doi: 10.3934/QFE.2017.1.26
[2]
Timotheos Paraskevopoulos, Peter N Posch .
A hybrid forecasting algorithm based on SVR and wavelet decomposition. Quantitative Finance and Economics, 2018, 2(3): 525-553.
doi: 10.3934/QFE.2018.3.525
[3]
Mustafa Hassan Mohammad Adam .
Nexus among foreign direct investment, financial development, and sustainable economic growth: Empirical aspects from Sudan. Quantitative Finance and Economics, 2022, 6(4): 640-657.
doi: 10.3934/QFE.2022028
[4]
Oleg Sukharev, Ekaterina Voronchikhina .
Financial and non-financial investments: comparative econometric analysis of the impact on economic dynamics. Quantitative Finance and Economics, 2020, 4(3): 382-411.
doi: 10.3934/QFE.2020018
[5]
Ngo Thai Hung .
Equity market integration of China and Southeast Asian countries: further evidence from MGARCH-ADCC and wavelet coherence analysis. Quantitative Finance and Economics, 2019, 3(2): 201-220.
doi: 10.3934/QFE.2019.2.201
[6]
Huseyin Tastan, Sercin Sahin .
Low-frequency relationship between money growth and inflation in Turkey. Quantitative Finance and Economics, 2020, 4(1): 91-120.
doi: 10.3934/QFE.2020005
[7]
Samuel Kwaku Agyei, Ahmed Bossman .
Investor sentiment and the interdependence structure of GIIPS stock market returns: A multiscale approach. Quantitative Finance and Economics, 2023, 7(1): 87-116.
doi: 10.3934/QFE.2023005
[8]
Andres Fernandez, Norman R. Swanson .
Further Evidence on the Usefulness of Real-Time Datasets for Economic Forecasting. Quantitative Finance and Economics, 2017, 1(1): 2-25.
doi: 10.3934/QFE.2017.1.2
[9]
Sudeshna Ghosh .
Asymmetric impact of COVID-19 induced uncertainty on inbound Chinese tourists in Australia: insights from nonlinear ARDL model. Quantitative Finance and Economics, 2020, 4(2): 343-364.
doi: 10.3934/QFE.2020016
[10]
Anouar Ben Mabrouk, Sabrine Arfaoui, Mohamed Essaied Hamrita .
Wavelet-based systematic risk estimation for GCC stock markets and impact of the embargo on the Qatar case. Quantitative Finance and Economics, 2023, 7(2): 287-336.
doi: 10.3934/QFE.2023015
Abstract
Obesity and type 2 and diabetes mellitus (T2D) are two dual epidemics whose shared genetic pathological mechanisms are still far from being fully understood. Therefore, this study is aimed at discovering key genes, molecular mechanisms, and new drug targets for obesity and T2D by analyzing the genome wide gene expression data with different computational biology approaches. In this study, the RNA-sequencing data of isolated primary human adipocytes from individuals who are lean, obese, and T2D was analyzed by an integrated framework consisting of gene expression, protein interaction network (PIN), tissue specificity, and druggability approaches. Our findings show a total of 1932 unique differentially expressed genes (DEGs) across the diabetes versus obese group comparison (p≤0.05). The PIN analysis of these 1932 DEGs identified 190 high centrality network (HCN) genes, which were annotated against 3367 GO terms and functional pathways, like response to insulin signaling, phosphorylation, lipid metabolism, glucose metabolism, etc. (p≤0.05). By applying additional PIN and topological parameters to 190 HCN genes, we further mapped 25 high confidence genes, functionally connected with diabetes and obesity traits. Interestingly, ERBB2, FN1, FYN, HSPA1A, HBA1, and ITGB1 genes were found to be tractable by small chemicals, antibodies, and/or enzyme molecules. In conclusion, our study highlights the potential of computational biology methods in correlating expression data to topological parameters, functional relationships, and druggability characteristics of the candidate genes involved in complex metabolic disorders with a common etiological basis.
1.
Introduction
In general, one of the criteria for measuring economic growth is the real GDP per capita (RGDPPC) of a country. The discussion on the nature of economic variability always remains a hot topic at the center of macroeconomic research. An important issue is whether the business cycle is exclusively transitory or whether it is considered mostly by permanent rather than temporary shifts. Albeit, there is a conventional wisdom of the business cycle that hypothesizes that GDP growth rate fluctuations are perpetual deviations from trend. It requires employing macroeconomic stabilization strategy encompassing both monetary and fiscal policies to manage with GDP disturbance as put forward by Keynes arguing neoclassical macroeconomic view. The RGDPPC is one of the most critical signs of the affluence of nations. Therefore, it is crucial to determine whether the per capita income follows a trend stationary or a difference stationary process1.
Lucas (1977) mentions with the question, "Why is it that, in capitalist economies, aggregate variables undergo repeated fluctuations about trend, all of essentially the same character?" Similarly, Nelson and Plosser (1982) on the investigation of the time-series properties of real output levels reveal that "we are unable to reject the hypothesis that these series are non-stationary stochastic processes with no tendency to return to a trend line. Based on these findings and an unobserved components model for output that decomposes fluctuations into a secular or growth component and a cyclical component, we infer that shocks to the former, which we associate with real disturbances, contribute substantially to the variation in observed output". After the influential work of Nelson and Plosser (1982), several other studies examine the potential time-series properties of crucial macroeconomic variables. Nelson and Plosser expound that a unit root in real output is uneven with the belief that business cycles are stationary variations all over a deterministic trend. Many prior studies including Perron (1988) endorse the findings of Nelson and Plosser (1982).
In a similar vein, Campbell and Mankiw (1987) summarize that as per the conventional view of the business cycle, fluctuations in output characterize transitory deviations from the trend. The empirical findings reveal a natural complement to standard tests of non-stationarity, while, using quarterly postwar US data. The study further adds that several business cycles' traditional theories sustain two simple ideas. First, fluctuations in output are supposed to be determined mainly by shocks to aggregate demand, like fiscal and monetary policies, or animal spirits. Second, shocks to aggregate demand are supposed to have merely a transitory influence on output; the economy returns to the natural rate in the long run. Li (2000) notes the rational implication of the government set off structural reform would certainly be of limited value if real output holds a unit root. Smyth and Inder (2004) using RGDPPC find a unit root for the majority of provinces of China. The study suggests that the government launched structural reform will be offset by other shocks, limiting its effects on the long-run growth track for all these provinces in China.
To examine empirically the time-series properties of output, cross-sectional and panel data analysis have been carried out by several existing studies for different countries using different sample periods. Where visible criticism of unit root tests has been observed, notably the Augmented Dickey-Fuller (ADF) test due to its low power against substantial, but stationary replacements with generally existing time lengths of data. While some researchers are in favor of the panel unit root techniques because of having more power as compared to unit root tests designed for univariate time series2. However, every country indeed has its own administrative, cultural, economic, political and social characteristics; therefore, time series analysis has its own undeniable features over the other one. Thus, the primary purpose of the present study is to empirically investigate the stationarity property of output fluctuations for a set of 177 countries from the whole World over the period ranging from 1960 to 2018. The critical feature of this study is that, to the best of the authors' knowledge, the research in hand is the first of its kind to employ the newly developed wavelet unit root tests proposed by Aydin and Pata (2020)3. These tests are developed in the wavelet framework that takes into account both frequency as well as time dimensions and also considers the issue of smooth structural break(s), thus it provides more robust inferences as compared to existing wavelet unit root tests such as Fan and Gencay (2010) which have low power particularly under negative moving averages and also when there are structural break(s) (see Eroglu and Soybilgen (2018) for more details). The wavelet tests proposed by Aydin and Pata (2020) are the wavelet version of popular Augmented Dickey Fuller (ADF) tests and named as Wavelet ADF (WADF) and its Fourier version—the Fourier WADF (FWADF) which considers the frequency dimension as well and which has the ability to take into account any smooth structural break(s). Thus, this study makes use of these recently developed WADF and FWADF by Aydin and Pata (2020). Further, this is the first study, using the maximum available numbers of countries along with a considerable length of the time period (minimum 25 years of annual observations), particularly, countries from all seven regions (East Asia and Pacific, Europe and Central Asia, Latin America and the Caribbean, Middle East and North Africa, North America, South Asia and Sub-Saharan Africa) across the globe are selected and analyzed. Therefore, the present study contributes significantly to the literature on the investigation of time series properties of per capita GDP for almost the entire World economies with more holistic empirical techniques.
Note that the the categorization of countries on the basis of geographical location is done to make the presentation of results more convenient and presentable. Previous, several studies on the energy-growth nexus, have adopted a similar approach and analyzed the nexus by considering different regions across the globe. See for example, Apergis and Payne (2009) for Central America; Eggoh et al. (2011) for African countries; Noor and Siddiqi (2010) for South Aisan countries; Jalil (2014) for net energy exporter and importer countries; Ozturk (2010) for a survey of studies on several geographical locations; Tiba and Omri (2017) for a survey of studies on energy, environment and economic growth; Rodríguez-Caballero and Ventosa-Santaulària (2017) for Latin American economies; Rodríguez-Caballero (2021) for four different regions, Asia and Pacific, Europe, Africa, The Americas. All these studies make use of conventional unit root tests (considering time dimension only) to assess the stationarity of the chosen data series as part of their empirical analysis. However, as highlighted in Aydin and Pata (2020) and the findings of current study show that ignoring the frequency dimension while applying unit root test may lead to misleading inferences regarding the stationarity of series under examination. So, one may re-assess the claims made by these studies by following our approach which provides better insights, as it not only takes into account the time dimension but also the frequency dimension as well along with (any) smooth structural change. Note that, the above studies are taken from energy-growth literature, however, there are several other studies, where real GDP per capita (RGDPPC) series is considered and its nexus is considered with other potential (macroeconomic) time series, and stationarity of RGDPPC is established using conventional (time domain based) unit root tests as part of their empirical analysis, so their results should also be re-analyzed and updated under the light of empirical findings presented in our study.
The remainder of the study is structured as follows. Section 2 deals with the literature review of the research. Section 3 introduces the econometric methodology used. Section 4 interprets the empirical findings while Section 5 concludes the study.
2.
Literature review
The available literature provides much empirical evidence on the investigation of time-series properties of macroeconomic variable namely per capita GDP or GDP or Gross National Product GNP (real). However, the empirical results of these previous studies are yet elusive. For example, Nelson and Plosser (1982) find that US real GDP to be a non-stationary process using annual data, mostly averages for the year, with a start from 1860–1909 and ending in 1970 in all cases. Campbell and Mankiw (1987) reveal regarding short-run dynamics that make GNP divergent from a random walk with drift. The long-run implications of the study estimates indicate that shocks to GNP are mostly permanent. Haan and Zelhorst (1993) observe that the traditional unit root tests fail to discard the null hypothesis of a unit root which exhibits stationarity of output in 12 OECD countries with the exception of US. Similarly, Li (2000) shows that China's output series is trend stationary while employing unit root tests with structural breaks for the empirical analysis carried over the period 1952 to 1998.
Chang et al. (2005) intensely reject the null of unit root process for one-third in 26 African countries during 1960–2000. Based on the empirical results, their study suggests that fiscal and/or monetary policies are expected to have only provisional impacts on the real output levels of over one-third out of 26 African countries. Narayan (2007) rejects the unit root null hypothesis for G7 countries at the 5% level except for Germany and Italy, over the period from 1870–2001, while employed Lagrange Multiplier test. In a similar vein, Narayan (2008) investigates whether GDP per capita for 15 Asian countries is panel stationary. The study indicates prodigious evidence of panel stationarity of real GDP per capita for different panels in the context of Asian countries. The empirical findings of Chen (2008) favor the view that the real GDP per capita of six countries namely Australia, France, Germany, Japan, UK and USA is illustrated by a stationary process when making use of one-break unit root test. However, the study rejects the null of unit root for 11 of 19 developed countries' real GDP per capita when employed the two-break unit root test. Thus, it suggests that results are reliable with the view that business cycles show stationary fluctuations around a deterministic trend. Chang et al. (2010) empirical results direly reject the unit root process for most of the Chinese provinces during 1952–1998 while the null doesn't get rejected for 11 out of 25 provinces.
Cuestas and Garratt (2011) observe evidence of a worldwide stationary ESTAR process around a nonlinear deterministic trend in more or less half of the studied countries. The study further adds that our findings reveal that nonlinearities act on real GDP series, while abandoning them; one can reach to ambiguous conclusions from unit root tests. Chang et al. (2011) confirm that real GDP is well exemplified by a non-linear mean returning process in three out of nine Central Eastern European (CEE) countries namely Bulgaria, Latvia and Romania during 1969–2009. Tiwari et al. (2012a) investigate stationarity properties of GDP per capita for 17 Asian countries4 and sub-panels during 1950–2009. The empirical results of the second-generation tests for unit root show stationarity of GDP per capita for the whole Asian panel and even for the East Asian and high-income Asian sub-panels as well. However, in the case of sub-panel for the South Asia, the study provides a weak indication for stationarity of GDP per capita. While, in another study by Tiwari et al. (2012b), the results show that real GDP per capita is nonlinear stationary indicating that shocks to economy by external or internal policies have lasting influence on real per capita GDP of SAARC countries during 1980–2010. Shen et al. (2013) find that a real shock would be lastingly affect the real per capita real GDP of three CEE countries namely, Bulgaria, Hungary and Poland during 1991–2012. The empirical evidence supports the real shocks which have a transitory influence on the long-run output levels, implying in these six CEE countries that their real GDP per capita settlement is a mean reversion in the direction of equilibrium values which shows periods of exploding behavior. The study further maintains that real GDP per capita follows a steady growth rate where policy implications then have transitory influences in the case of 6 CEE countries (i.e., Czech Republic, Estonia, Latvia, Lithuania, Romania, and Slovakia).
4Countries includes Bangladesh, Cambodia, China, Hong Kong, India, Indonesia, Japan, Malaysia, Myanmar, Pakistan, Philippines, Singapore, South Korea, Sri Lanka, Taiwan, Thailand, and Vietnam.
The findings of Jannati et al. (2013) reveal that the real GDP per capita over the period 1970–2009 in 5 countries is stationary while carrying out empirical analysis of 16 countries in Asia with non-linear approaches to test for stationarity. Dogru (2014) evaluates the stationarity property of real GDP per capita fluctuations for a panel of 11 MENA countries during 1970–2012. The empirical results of both seemingly unrealated ADF (SURADF) and cross-sectional ADF (CADF) approaches lead support to stationarity of real GDP per capita. The study demonstrates that the shocks to the output are mostly lasting rather than temporary. The empirical findings of Chang et al. (2014) from numerous conventional unit root tests validate that real GDP per capita for 52 African countries is non-stationary, while, when the SPSM using the Panel KSS unit root test with a Fourier function is employed, the study observes that real GDP per capita are stationary in 50 out of 52 selected countries. The empirical outcomes suggest that real GDP per capita fluctuations for these countries are inclusively transitory during 1969v2011. Using the BCIPS panel unit root test by Zeren and İşlek (2019) in the case of D8 economies from 1960–2014 reveal that GDP per capita is stationary in multifactor situations in chosen D8 economies. Using Box-Jenkins approach (known as ARIMA), Nyoni and Muchingami (2019) suggest that Botswana GDP per capita is not mean reverting over the period 1960–2017 with stable estimates, indicating that in next few years, the overall standard of life will for Botswanian people will improve.
Table 1 portrays the summary of erstwhile empirical work available analyzing the stationarity property of output fluctuations. The forgoing analysis demonstrates that evidently the empirical results are inconclusive and still the stationarity property of output fluctuations is unsolved and further research on the topic under the study is required.
Table 1.
Previous studies on stationarity properties of output fluctuations.
This paper makes use of recently developed unit root tests developed by Aydin and Pata (2020) based on a general wavelet spectral approach. Working in parallel to Aydin and Pata (2020) and Yazgan and Ozkan (2015), the following DGP is considered which takes into account the smooth structural change as well: xt=γ(t)+ϵt
where, γ(t) is unknown deterministic component (Becker et al., 2006), given by:
γ(t)=β∑ni=1[(2i−1)−1sin{(2π(2i−1)kt)/T}]
(1)
Note that "k" and "n" respectively denotes frequency of fourier term and deterministic component and when n = 1, the series contains a smooth break and when n > 1, break becomes more sharp and abrupt (see Aydin and Pata (2020) and Yazgan and Ozkan (2015) for more details). The case of smooth break (n = 1) is considered by Aydin and Pata (2020) by using the Wavelet ADF (WADF) test by Eroglu and Soybiligen (2018). The same is considered in this study and is given below:
∆Y1,t=∑pi=1θi∆Y1,t−i+τY1,t−1+δsin(2πktT)+ϵt
(2)
where, Y1,t indicates scaling coefficients that show low frequency components obtained via wavelet filter (we used Haar filter, which according to Aydin and Pata (2020) is powerful). The focus is made on testing the following null and alternative hypotheses:
H0: τ = 0 (unit root) against H1: τ < 0 (no unit root)
The null is tested via standard ADF statistic:
ADF∗t=(ˆτ−τH0)/SE(ˆτ)
(3)
where, ˆτ and SE(ˆτ) are the estimate of τ and its corresponding standard error.
To carryout FWADF unit root test, a two-step procedure suggested in Enders and Lee (2012) is adopted, where, model in Equation (2) is estimated for different frequencies, k = 1, 2, …, 5., and residuals sum of squares (RSS) is calculated for each model and then a model is selected among all these candidate models which RSS. In the next step, a standard t-test is applied to has lowest value of assess non-linearity and if fourier term in Equation (2) is found to be insignificant then WADF unit root test is applied, and in case of significant fourier term, the FWADF unit root test is applied. The critical values are obtained by following guidelines in Becker et al. (2006) and the same are provided in Table 10 (A–C) in Supplementary for different sample sizes that we considered in this study.
3.2. Data and its sources
We use annual time series data on real GDP per capita of all countries across the world. It is measured in constant 2010 US dollars. The data source is WDI 2020, the World Bank database. The total number of countries selected for the analysis are 177 out of a total available 218 countries on WDI. For the remaining 41 countries, either the data was not available at WDI or the data span was less than 25 years. We set minimum limit to 25 data points to get reasonable power of the wavelet tests considered in this study even under small samples. The countries are classified into seven different regions (East Asia and Pacific Region, Europe and Central Asia, Latin America and the Caribbean, Middle East and North Africa, North America, South Asia and Sub-Saharan Africa) as per guidelines of World Bank. Note that the idea of classifying countries is not necessary, however, this is done to present the empirical results in a more meaningful way as most of the time, the empirical researchers are interested in carrying out analysis by considering a particular region. The same countries can be sub categorized as per different income groups as well, but we thought it better to present the analysis region-wise as this will give us deep insights of the issue at hand and it will portray a clear picture for different regions of the world which is currently not available anywhere in the existing literature. The time period for most of the countries is 1960 to 2018, while for some countries, we use a different time period to make use of maximum available data for the analysis of chosen countries. The names of countries along with their three-digit code as per WDI guidelines and the selected time period is provided in each table for clarity. The results of basic summary statistics are calculated but are not reported here to save space but these are available from authors upon request.
4.
Empirical findings and discussions
This section provides results of wavelet-based unit root tests—the WADF and FWADF. Please note that as per World Bank, the total number of countries included in East Asia and Pacific Region are 38. Among these 38 countries, we are able to get data of 26 of them with at-least 25 observations, and carried out analysis for these 26 countries reported in Table 2. The empirical findings reveal useful insights of the behavior of the time series under consideration. Specifically, only one country, Brunei Darussalam is found to be stationary under both specifications (constant as well as constant and trend) while for 4 out of 26 countries (Brunei Darussalam, Vanuatu, Samoa and Cambodia), the null of non-stationarity gets rejected under constant and trend specification and for 3 out of 26 countries (Japan, Micronesia Fed. Sts. and Brunei Darussalam), null gets rejected under only constant specification. Overall, for 6 countries (Japan, Micronesia Fed. Sts., Vanuatu, Samoa, Cambodia and Brunei Darussalam), the null of unit root gets rejected under any of the two specifications (with constant and constant plus trend) considered. So in total, out of 26 countries, the null hypothesis that the series has a unit root gets rejected by 6 countries while for the rest of 20 countries, the series is found to be stationary and null couldn't get rejected by the wavelet tests considered.
Table 2.
East Asia and Pacific Region.
S. No.
Country
Code
Time span
sample size
With Constant
With Constant and Trend
k
p
FWADF
t-stat
p
WADF
k
p
FWADF
t-stat
p
WADF
1
Australia
AUS
1960–2018
59
1
0
−1.13
−2.55***
-
-
1
0
0.28
−2.13**
-
-
2
China
CHN
1960–2018
59
4
1
0.17
2.17
1
0.24
4
1
0.09
1.87
1
0.14
3
Fiji
FJI
1960–2018
59
5
1
0.20
−1.94**
-
-
1
1
−4.07
3.13
1
−2.51
4
Indonesia
IDN
1960–2018
59
3
0
3.07
−2.20**
-
-
3
0
−0.37
−2.20**
-
-
5
Japan
JPN
1960–2018
59
5
0
−2.72*
−1.72*
-
-
5
0
−0.41
−1.68*
-
-
6
Korea, Rep.
KOR
1960–2018
59
1
0
0.86
−3.46***
-
-
1
0
−2.02
−2.92***
-
-
7
Myanmar
MMR
1960–2018
59
1
1
0.12
−2.25**
-
-
5
1
−1.62
−2.00**
-
-
8
Malaysia
MYS
1960–2018
59
3
0
3.12
−1.98**
-
-
3
0
−0.32
−1.76*
-
-
9
Philippines
PHL
1960–2018
59
3
1
1.38
−1.36
1
1.40
1
1
−3.97
4.79
1
−0.01
10
Papua New Guinea
PNG
1960–2018
59
3
0
−1.37
−2.75***
-
-
1
1
−3.29
2.40
1
−2.22
11
Singapore
SGP
1960–2018
59
3
0
2.20
−0.90
0
2.46
3
0
−1.73
−2.16**
-
-
12
Thailand
THA
1960–2018
59
4
0
3.02
2.61
0
2.41
4
0
−1.34
2.60
0
−1.49
13
Hong Kong SAR, China
HKG
1961–2018
58
4
0
2.28
2.06
0
1.93
4
0
−2.06
2.38
0
−1.83
14
Kiribati
KIR
1970–2018
49
1
2
−8.38
1.56
3
−3.75***
1
2
−6.97
1.48
3
−2.46
15
New Zealand
NZL
1970–2018
49
1
0
−0.22
−1.57*
-
-
5
1
−2.67
−1.75*
-
-
16
Brunei Darussalam
BRN
1974–2018
45
1
2
−2.81
−1.55*
-
-
1
2
−4.52**
−2.63**
-
-
17
Vanuatu
VUT
1979–2018
40
3
2
0.00
3.24
2
-0.56
4
2
−3.45
1.71
3
−5.08***
18
Mongolia
MNG
1981–2018
38
1
0
1.01
−2.76***
-
-
5
2
−0.71
2.34
0
−0.28
19
Tonga
TON
1981–2018
38
5
0
−0.28
−1.16
0
−0.37
2
0
−4.04
2.37
0
−2.42
20
Macao SAR, China
MAC
1982–2018
37
5
1
−1.30
1.87
1
−1.32
2
1
−1.50
−1.81**
-
-
21
Samoa
WSM
1982–2018
37
2
1
0.56
2.96
0
−0.07
1
2
−4.17
1.32
2
−3.82**
22
Lao PDR
LAO
1984–2018
35
1
2
1.88
−1.29*
2
1.66
2
0
10.94
−10.7***
-
-
23
Vietnam
VNM
1984–2018
35
3
1
1.90
0.53
1
2.24
3
0
0.34
4.71
0
−0.48
24
Micronesia, Fed. Sts.
FSM
1986–2018
33
1
0
−4.42**
−3.27***
-
-
1
0
−3.69
−3.95***
-
-
25
Tuvalu
TUV
1990–2018
29
3
0
−1.61
1.53
0
−1.44
3
0
−2.83
1.24
0
−2.62
26
Cambodia
KHM
1993–2018
26
3
1
0.99
−1.70**
-
-
3
0
−5.13***
−2.40**
-
-
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
The results related to countries included in Europe and Central Asia are provided in Table 3. According to guidelines by World Bank, the total number of categories in this region are 58. From these 58 countries, we are able to get data of 46 of them with at-least 25 observations, and carried out analysis for these 46 countries. The empirical findings reveal useful insights of the behavior of the time series under consideration. Specifically, in case of model with constant specification, 7 out of 46 countries namely, France, Greece, Georgia, Tajikistan, Ukraine, Russian Federation and Belarus are found to be stationary while rest of 39 countries have non-stationary per capita GDP, and when the constant and trend specification is considered, 21 countries out of 46, namely, Spain, Finland, Greece, Iceland, Norway, Andorra, Albania, Tajikistan, Kyrgyz Republic, Ukraine, Azerbaijan, Czech Republic, Kazakhstan, North Macedonia, Romania, Slovenia, Hungary, Slovak Republic, Bosnia and Herzegovina and Serbia are found to be stationary while rest of 25 countries have non-stationary per capita GDP and for three countries (Greece, Tajikistan and Ukraine), the null of unit root gets rejected under both specifications considered (constant and constant plus trend), while for 25 countries the null gets rejected under any of the two specifications considered. So in total, out of 46 countries, the null hypothesis that the series has a unit root gets rejected by 25 countries while for the rest of 21 countries, the series is found to be stationary and null couldn't get rejected by any of the specifications considered.
Table 3.
Europe and Central Asia.
S. No.
Country
Code
Time span
sample size
With Constant
With Constant and Trend
k
p
FWKSS
t-stat
p
WADF
k
p
FWKSS
t-stat
p
WADF
1
Austria
AUT
1960–2018
59
1
0
−2.39
−1.79*
-
-
1
1
−1.76
−1.89**
-
-
2
Belgium
BEL
1960–2018
59
2
0
−0.81
1.56
0
−1.55
1
1
−2.38
−1.36
1
−2.18
3
Denmark
DNK
1960–2018
59
2
0
−0.60
1.75
0
−1.31
1
0
−3.32
−2.42**
-
-
4
Spain
ESP
1960–2018
59
2
2
−0.58
2.16
2
−1.11
3
1
−4.10
0.89
1
-4.16***
5
Finland
FIN
1960–2018
59
3
0
−0.25
2.52
0
−0.72
3
1
−3.12
1.71
1
-3.47**
6
France
FRA
1960–2018
59
2
0
−2.13
1.51
0
−2.94**
2
0
0.02
1.50
0
−0.52
7
United Kingdom
GBR
1960–2018
59
1
0
−1.53
−1.96**
-
-
1
0
−2.56
−2.28**
-
-
8
Greece
GRC
1960–2018
59
1
1
−3.05*
−1.95*
-
-
3
1
−3.09
1.28
1
-3.17*
9
Iceland
ISL
1960–2018
59
3
0
0.21
1.66
0
−0.12
2
1
−4.70***
−2.68***
-
-
10
Italy
ITA
1960–2018
59
2
1
−1.70
1.24
1
−1.94
2
0
1.04
1.37
0
1.06
11
Luxembourg
LUX
1960–2018
59
1
0
−2.21
−2.89***
-
-
1
1
−3.08
−3.25***
-
-
12
Netherlands
NLD
1960–2018
59
1
1
−2.16
−2.18**
-
-
5
1
−2.67
1.10
1
−2.70
13
Norway
NOR
1960–2018
59
1
1
−1.92
−1.54*
1
−1.17
1
1
−3.47*
−3.18***
-
-
14
Portugal
PRT
1960–2018
59
2
2
−1.53
2.85
2
−1.72
2
1
0.34
2.78
1
-2.57
15
Sweden
SWE
1960–2018
59
3
0
0.81
2.55
0
0.17
3
0
−1.20
2.30
0
-1.62
16
Turkey
TUR
1960–2018
59
3
0
2.97
−0.67
0
3.30
3
0
2.08
−1.56*
-
-
17
Georgia
GEO
1965–2018
54
2
1
−3.03
−1.52*
1
−2.65*
1
1
−4.38
3.13
1
−2.57
18
Andorra
AND
1970–2018
49
1
1
−1.66
−1.85**
-
-
2
1
−1.15
1.95
1
−3.09*
19
Germany
DEU
1970–2018
49
4
0
−0.56
−1.39
0
−0.42
4
0
−3.11*
−1.58*
-
-
20
Greenland
GRL
1970–2018
49
1
0
−0.75
−1.28
0
0.01
1
1
−2.19
1.41
1
−1.86
21
Ireland
IRL
1970–2018
49
4
1
0.71
−1.38
3
1.33
2
1
−2.67
1.70
1
−2.49
22
Monaco
MCO
1970–2018
49
4
2
2.44
−1.84**
-
-
4
2
0.59
−1.80**
-
-
23
Cyprus
CYP
1975–2018
44
2
0
−1.69
1.97
0
−1.56
1
1
−2.35
−1.41*
-
-
24
Albania
ALB
1980–2018
39
1
1
−0.49
−4.54***
-
-
4
1
−2.21
−1.21
3
−3.79**
25
Bulgaria
BGR
1980–2018
39
2
0
3.12
3.55
0
1.23
4
1
−1.61
−2.49**
-
-
26
Isle of Man
IMN
1984–2017
34
4
0
1.67
−2.51**
-
-
2
2
−2.25
1.95
2
-2.61
27
Tajikistan
TJK
1985–2018
34
2
1
−3.35**
−2.14**
-
-
2
2
−4.90
0.97
3
−10.35***
28
Kyrgyz Republic
KGZ
1986–2018
33
1
1
−0.37
−3.23***
-
-
4
2
−5.67
−0.91
3
−9.98***
29
Turkmenistan
TKM
1987–2018
32
1
1
3.00
−4.31***
-
-
5
2
−2.96
2.84
2
−1.91
30
Ukraine
UKR
1887–2018
32
1
1
−3.89***
−2.29**
-
-
5
2
−5.31
1.67
3
−5.08***
31
Uzbekistan
UZB
1987–2018
32
1
2
4.39
−6.00***
-
-
1
2
0.51
−3.48***
-
-
32
Russian Federation
RUS
1989–2018
30
5
2
−2.09
1.53
3
-2.97**
5
2
−3.25
2.16
3
−2.92
33
Armenia
ARM
1990–2018
29
3
1
−2.14
−3.07***
-
-
3
2
−1.51
−1.99**
-
-
34
Azerbaijan
AZE
1990–2018
29
1
1
−3.24
−4.00***
-
-
1
1
−4.60***
−4.27***
-
-
35
Belarus
BLR
1990–2018
29
1
2
−10.15***
−9.70***
-
-
1
1
−2.25
−10.7***
-
-
36
Czech Republic
CZE
1990–2018
29
2
1
−0.52
3.95
1
−0.49
2
2
−3.89
3.81
3
−3.97**
37
Kazakhstan
KAZ
1990–2018
29
5
2
−2.38
1.82
2
−2.22
1
1
−5.00**
−3.00***
-
-
38
North Macedonia
MKD
1990–2018
29
1
1
0.35
−1.96**
-
-
1
0
−2.14
−1.31*
0
−5.57***
39
Romania
ROU
1990–2018
29
2
0
3.32
5.05
1
0.48
2
2
0.19
2.63
3
−3.17*
40
Slovenia
SVN
1990–2018
29
2
1
−2.09
1.73
3
−1.74
1
1
−6.36***
−4.68***
-
-
41
Hungary
HUN
1991–2018
28
2
1
−0.25
−1.57**
1
−0.39
2
1
−4.46
2.48
1
−3.52**
42
Slovak Republic
SVK
1992–2018
27
2
0
0.88
4.94
0
0.20
3
2
−5.01***
−2.62**
-
-
43
Estonia
EST
1993–2018
26
4
0
−0.83
1.45
0
−0.76
2
0
−3.39
2.59
0
−1.97
44
Bosnia and Herzegovina
BIH
1994–2018
25
1
2
−2.29
3.30
2
−2.28
2
2
−5.44
2.75
3
−5.91***
45
Latvia
LVA
1994–2018
25
2
0
−1.76
1.87
0
−0.56
2
0
−4.69
4.77
1
−2.33
46
Serbia
SRB
1994–2018
25
5
2
−1.47
−1.52**
2
−1.66
4
2
−4.76
2.23
2
−3.34*
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
The empirial results for the Latin America and the Caribbean region are presented in Table 4 and as per World Bank, this region constitutes of a total of 42 countries. We were lucky to obtain data on 36 countries comprising of at-least 25 annual observations and thus, the empirical analysis is done for the 36 countries. The empirical findings provide detailed analysis of data series considered. Specifically, considering the model with constant specification, 7 out of 36 countries, namely, Bahamas, Suriname, Venezuela RB, Jamaica, Barbados, Dominica and Aruba are found to have a stationary per capita GDP, while rest of 29 countries have non-stationary per capita GDP, and in case of model with constant and trend specification, 12 countries out of 36, namely, Bahamas, Belize, Brazil, Ecuador, Guatemala, Suriname, Trinidad and Tobago, Venezuela RB, Jamaica, Switzerland, Grenada and Aruba are found to have stationary per capita GDP, while rest of the 19 countries have non-stationary per capita GDP. In addition, 5 countries (Bahamas, Suriname, Venezuela RB, Jamaica and Aruba) have stationary per capita GDP under both specifications, while 14 out of selected 46, have per capita GDP stationary under either of the two specifications. These include, Bahamas, Belize, Brazil, Ecuador, Guatemala, Suriname, Trinidad and Tobago, Venezuela RB, Jamaica, Switzerland, Barbados, Dominica, Grenada and Aruba. So in total, out of 36 countries, the null hypothesis that the series has a unit root gets rejected by 14 countries while for the rest of 22 countries the series is found to be non-stationary and null couldn't get rejected by any of the tests considered.
Table 4.
Latin America and The Caribbean.
S. No.
Country
Code
Time span
sample size
With Constant
With Constant and Trend
k
p
FWKSS
t-stat
P
WADF
k
p
FWKSS
t-stat
P
WADF
1
Argentina
ARG
1960–2018
59
4
0
−1.12
1.07
0
−0.67
1
0
−3.44
2.33
1
−2.55
2
Bahamas
BHS
1960–2018
59
1
0
−4.46***
−3.41***
-
-
1
1
−4.26**
−2.64***
-
-
3
Belize
BLZ
1960–2018
59
1
1
−2.90
−2.83***
-
-
1
1
−4.21**
−3.05***
-
-
4
Bolivia
BOL
1960–2018
59
3
1
−0.65
−2.22***
-
-
1
1
−5.32
4.61
1
−2.47
5
Brazil
BRA
1960–2018
59
5
1
−1.73
1.78
1
−1.72
2
1
−3.97**
−2.81***
-
-
6
Chile
CHL
1960–2018
59
1
0
−0.21
−3.28***
-
-
1
0
−0.29
−2.29**
-
-
7
Colombia
COL
1960–2018
59
3
2
1.95
−2.64***
-
-
1
1
−3.70
2.61
1
−2.37
8
Costa Rica
CRI
1960–2018
59
3
0
1.89
−2.02**
-
-
3
0
−0.10
−2.03***
-
-
9
Dominican Republic
DOM
1960–2018
59
5
0
4.34
1.86
0
3.99
5
0
1.60
1.87
0
1.31
10
Ecuador
ECU
1960–2018
59
2
1
−1.30
−2.23**
-
-
2
1
−4.08**
−2.63***
-
-
11
Guatemala
GTM
1960–2018
59
3
1
−1.14
−2.29**
-
-
1
1
−3.82
2.15
1
−2.95*
12
Guyana
GUY
1960–2018
59
3
1
2.15
−3.28***
-
-
3
1
0.35
−3.06***
-
-
13
Honduras
HND
1960–2018
59
4
0
0.33
1.40
0
0.10
1
0
−5.98
5.41
0
−1.32
14
Haiti
HTI
1960–2018
59
1
1
−2.96
2.71
1
−1.35
2
1
−3.20
−3.66***
-
-
15
Mexico
MEX
1960–2018
59
5
0
−1.80
−1.90**
-
-
5
0
−1.83
−1.68*
-
-
16
Nicaragua
NIC
1960–2018
59
1
1
−2.02
1.46
1
−1.68
1
1
−2.51
2.04
1
−1.45
17
Panama
PAN
1960–2018
59
3
0
3.15
−0.62
0
3.45
1
0
−1.17
1.48
0
1.39
18
Peru
PER
1960–2018
59
1
1
−0.30
−2.32**
-
-
4
1
−1.60
−1.03
1
−0.43
19
Puerto Rico
PRI
1960–2018
59
1
0
−2.57
−2.27**
-
-
1
0
−1.75
−2.31**
-
-
20
Paraguay
PRY
1960–2018
59
4
1
−0.08
1.88
1
−0.31
4
1
−1.94
1.33
1
−2.44
21
Suriname
SUR
1960–2018
59
2
1
−4.04***
−2.58***
-
-
2
1
−5.17***
−3.62***
-
-
22
Trinidad and Tobago
TTO
1960–2018
59
1
1
−2.79
−2.26**
-
-
5
1
−3.15*
−2.07**
-
-
23
Uruguay
URY
1960–2018
59
3
0
1.55
−3.06***
-
-
3
1
−0.68
−2.96***
-
-
24
St. Vincent and the Grenadines
VCT
1960–2018
59
1
0
−1.37
−2.63***
-
-
1
0
−2.69
−2.06**
-
-
25
Venezuela, RB
VEN
1960–2018
59
1
1
−3.21
1.94
1
−2.44*
2
1
−3.85**
−2.06**
-
-
26
El Salvador
SLV
1965–2018
54
4
1
−1.05
−1.45
1
−1.10
4
1
−1.71
−1.42
1
−1.77
27
Jamaica
JAM
1966–2018
53
1
1
−3.72**
−2.43***
-
-
2
1
−4.14
2.25
3
−4.81***
28
Switzerland
CHE
1970–2018
49
3
0
0.36
2.19
0
−0.16
2
1
−4.87**
−1.91**
-
-
29
Cuba
CUB
1970–2018
49
4
1
−0.78
1.23
1
−0.95
1
1
−3.20
2.37
1
−2.04
30
Barbados
BRB
1974–2018
45
1
0
−2.30*
−1.85**
-
-
5
1
−2.21
−2.32**
-
-
31
Antigua and Barbuda
ATG
1997–2018
42
5
0
−1.86
−1.43
0
−1.85
5
0
−1.80
−1.35*
0
−1.87
32
Dominica
DMA
1977–2018
42
1
1
−3.67**
−2.40**
-
-
1
1
−1.09
−2.35**
-
-
33
Grenada
GRD
1977–2018
42
2
0
0.18
1.60
2
−0.16
2
1
−4.31
1.15
1
−4.35**
34
St. Kitts and Nevis
KNA
1977–2018
42
3
0
−1.80
−0.99
0
−1.73
3
0
−0.21
−0.93
0
−0.62
35
St. Lucia
LCA
1977–2018
42
2
0
−2.57
−1.74*
-
-
3
2
−2.68
1.00
3
−2.52
36
Aruba
ABW
1986–2017
32
5
1
−1.66
1.60
0
−6.00***
4
0
−7.33***
−2.30***
-
-
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
Table 5 provides the findings of unit root tests for the Middle East and North African region which comprises of 21 countries in total (as per World Bank). A careful inspection yields data on 17 out of 21 countries with a minimum of 25 years of span. Thus, the sample for this group comprises of 17 countries. The empirical results are interesting in the sense that for the model with constant specification, 7 out of 17 countries, namely, Iran, Islamic Rep., Oman, Saudi Arabia, United Arab Emirates, Jordan, Bahrain and Lebanon are found to have a stationary series for per capita GDP, while the rest of 10 countries have non-stationary per capita GDP, similarly, for the model with constant and trend specification, 9 countries out of 17 considered, namely, Algeria, Egypt, Israel, Malta, United Arab Emirates, Jordan, Bahrain, Lebanon and West Bank & Gaza are found to be stationary while rest of 8 countries have non-stationary per capita GDP. In addition, 4 countries (United Arab Emirates, Jordan, Bahrain and Lebanon) have stationary per capita GDP under both specifications, while 12 out of selected 17, have per capita GDP stationary under either of the two specifications. So in total, out of 17 countries, the null hypothesis that the series has a unit root gets rejected by 12 countries while for the rest of 5 countries, the series is found to be non-stationary and null couldn't get rejected under any of the two specifications considered.
Table 5.
Middle East and North Africa.
S. No.
Country
Code
Time span
sample size
With Constant
With Constant and Trend
k
P
FWKSS
t-stat
P
WADF
k
p
FWKSS
t-stat
p
WADF
1
Algeria
DZA
1960–2018
59
1
2
−2.09
−1.31
2
−1.60
2
2
−6.60***
−6.31***
-
-
2
Egypt, Arab Rep.
EGY
1960–2018
59
2
2
1.80
−2.47***
-
-
2
2
−2.69*
−2.39**
-
-
3
Israel
ISR
1960–2018
59
3
0
−0.83
−2.34**
-
-
3
0
−3.20*
−2.99***
-
-
4
Iran, Islamic Rep.
IRN
1960–2017
58
1
1
−2.76
1.10
1
-2.52*
1
1
−3.07
1.72
1
−2.49
5
Oman
OMN
1965–2018
54
1
0
−4.39***
−2.57***
-
-
1
0
−2.80
−2.49**
-
-
6
Tunisia
TUN
1965–2018
54
2
0
0.58
1.54
0
0.81
2
1
−0.73
0.81
1
−1.55
7
Morocco
MAR
1966–2018
53
1
0
0.27
−1.25
1
1.33
2
0
−1.51
−1.15
1
−1.36
8
Iraq
IRQ
1968–2018
51
5
0
−0.27
1.88
0
−0.35
1
0
−3.60
2.44
0
−2.14
9
Saudi Arabia
SAU
1986–2018
51
4
2
−3.12**
−1.94**
-
-
4
2
−2.85
−1.79**
-
-
10
Malta
MLT
1970–2018
49
2
2
1.23
2.13
2
0.59
5
1
−3.74**
−1.83**
-
-
11
United Arab Emirates
ARE
1975–2018
44
2
2
−3.21
1.10
2
-3.00**
2
2
−3.90
1.84
2
−3.23*
12
Jordan
JOR
1975–2018
44
1
1
−3.14*
−1.81**
-
-
4
1
−2.69
−0.76
1
−3.13*
13
Bahrain
BHR
1980–2018
39
5
2
−3.86***
−2.63**
-
-
5
2
−3.46**
−2.48**
-
-
14
Lebanon
LBN
1988–2018
31
1
1
−3.92**
−2.17**
-
-
3
1
−3.77
2.99
3
−3.38*
15
Yemen, Rep.
YEM
1990–2018
29
4
1
−1.70
−1.36*
-
-
4
2
2.33
−1.87**
-
-
16
Kuwait
KWT
1992–2018
27
3
1
−2.13
2.02
1
−2.05
1
1
−3.52
−2.95***
-
-
17
West Bank and Gaza
PSE
1994–2018
25
1
0
1.01
−3.93***
-
-
3
2
−1.95
1.11
3
−3.26*
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
The North American region includes only three countries and all three have been considered in this study due to availability of relevant data. The results for these countries are provided in Table 6. The results suggest that for the model with constant specification, none of the three selected countries has stationary per capita GDP i.e., all countries have non-stationary per capita GDP, while, for the model with constant and trend specification, only one country (Bermuda) out of three, has stationary per capita GDP while the other two Canada and United States have non-stationary per capita GDP and none of the selected countries is found to be stationary under both specifications (constant as well as constant and trend), while only one country (Bermuda) is considered stationary under either of the two specifications. So for all the countries included in North America region, the GDP per capita series is found to be trend stationary and its in case of Bermuda only.
Table 6.
North America.
S. No.
Country
Code
Time span
samplesize
With Constant
With Constant and Trend
k
P
FWKSS
t-stat
p
WADF-stat
k
p
FWKSS
t-stat
p
WADF
1
Bermuda
BMU
1960–2013
54
1
0
−2.90
−2.07**
-
-
2
1
−4.22*
−3.19***
-
-
2
Canada
CAN
1960–2018
59
3
0
0.24
1.69
0
−0.18
2
0
−2.63
−1.98**
-
-
3
United States
USA
1960–2018
59
3
0
0.13
1.57
0
−0.18
3
0
−2.44
2.00
0
−2.14
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
The total number of countries in South Asia is eight. A total of 6 countries were able to meet the goal of containing a minimum of 25 years of data, so these countries are sorted for this region and final analysis is carried out for these and is presented in Table 7. The empirical findings suggest that under constant specification, the per capita GDP of all 6 selected countries is found to be stationary, while, under constant and trend specification, the per capita GDP of 5 out of 6 selected countries is found to be stationary. In addition, none of the selected countries is found to be stationary under both specifications while only one is considered stationary under either of the two specifications. Thus, out of 6 countries, the null hypothesis that the series has a unit root gets rejected by only one country, while for the rest of 5 countries the series is found to be stationary and null couldn't be rejected under any of the two specifications considered.
Table 7.
South Asia.
S. No.
Country
Code
Time span
sample size
With Constant
With Constant and Trend
K
p
FWKSS
t-stat
p
WADF
k
p
FWKSS
t-stat
p
WADF
1
Bangladesh
BGD
1960–2018
59
4
0
9.11
1.46
0
8.82
4
0
4.10
1.38
0
3.85
2
Bhutan
BTN
1980–2018
39
1
0
3.48
−1.55*
0
5.65
5
0
0.63
1.31
0
0.45
3
India
IND
1960–2018
59
5
2
6.31
1.35
2
6.08
5
2
5.81
1.30
2
5.58
4
Sri Lanka
LKA
1961–2018
58
2
0
6.56
−1.49*
0
7.41
2
0
1.50
−1.35
0
1.53
5
Nepal
NPL
1960–2018
59
3
0
7.58
−1.20
0
7.97
3
0
2.08
−1.17
0
2.14
6
Pakistan
PAK
1960–2018
59
5
1
0.43
−1.62*
-
-
2
1
−3.93
1.84
1
−3.37**
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
There are 48 countries in the Sub-Saharan Africa region (as per World Bank). A sample of 43 is finalized after removing the countries that had fewer number of observations, fewer than 25 (the set limit). The empirical results presented in Table 8 reveal that for the model under constant specification, 6 out of 43 countries, namely, Niger, Togo, Zambia, Zimbabwe, Guinea-Bissau and Equatorial Guinea are found to be stationary while rest of 37 countries have non-stationary per capita GDP, while for the model with constant and trend specification, 7 countries out of 43, namely, Botswana, Central African Republic, Cameroon, Zimbabwe, Gambia, Equatorial Guinea and Tanzania are found to be stationary while rest of 36 countries have non-stationary per capita GDP. In addition, 2 countries (Zimbabwe and Equatorial Guinea) have stationary per capita GDP under both specifications, while 11 out of selected 43 countries, have per capita GDP stationary under either of the two specifications. Overall, out of 43 selected countries, the null hypothesis of unit root gets rejected by 11 countries while for the remaining 32 countries we find evidence in favor of null under any of the two specifications considered.
Table 8.
Sub-Saharan Africa.
S. No.
Country
Code
Time span
sample size
With Constant
With Constant and Trend
k
p
FWKSS
t-stat
p
WADF
k
p
FWKSS
t-stat
p
WADF
1
Burundi
BDI
1960–2018
59
1
0
−1.46
2.68
0
−1.41
2
0
−1.53
−2.71***
-
-
2
Benin
BEN
1960−2018
59
4
0
1.95
−2.21**
-
−
4
0
−1.37
−1.68*
-
-
3
Burkina Faso
BFA
1960−2018
59
1
0
2.20
−1.24
0
3.86
3
0
0.35
1.09
0
0.30
4
Botswana
BWA
1960−2018
59
2
0
1.00
−1.70*
-
−
3
0
−3.08
−1.12
0
−3.11*
5
Central African Republic
CAF
1960−2018
59
1
0
−2.57
2.41
0
−0.68
2
1
−5.37***
−3.81***
-
-
6
Cote d'Ivoire
CIV
1960−2018
59
1
1
−4.25
3.65
1
−1.84
1
1
−4.15
3.03
1
−2.56
7
Cameroon
CMR
1960−2018
59
5
2
−2.93
1.26
1
−2.16
5
2
−3.30
1.32
2
−3.20*
8
Congo, Dem. Rep.
COD
1960−2018
59
4
1
−1.51
−3.29***
-
-
1
2
−5.05
4.08
2
−2.27
9
Congo, Rep.
COG
1960−2018
59
2
1
−2.33
−2.82**
-
-
2
1
−0.67
−3.01***
-
-
10
Gabon
GAB
1960−2018
59
3
0
−2.37
−1.69*
-
-
3
0
−2.00
−1.91**
-
-
11
Ghana
GHA
1960−2018
59
1
0
2.18
−3.38***
-
-
3
0
0.39
−1.25
0
0.61
12
Kenya
KEN
1960−2018
59
3
0
−1.12
−1.54*
0
−0.68
1
0
−3.92
4.73
0
−1.35
13
Lesotho
LSO
1960−2018
59
4
0
3.26
2.45
0
2.60
1
0
−1.45
1.02
0
−0.10
14
Madagascar
MDG
1960−2018
59
2
0
−1.49
1.48
0
−1.21
2
0
−0.01
1.50
0
−0.76
15
Mauritania
MRT
1960−2018
59
2
2
−1.32
−0.73
2
−1.12
1
2
−1.86
1.49
3
−1.93
16
Malawi
MWI
1960−2018
59
3
0
−0.98
−2.12**
-
-
1
0
−4.17
3.37
0
−1.37
17
Niger
NER
1960−2018
59
5
2
−3.24**
−2.33**
-
-
5
1
−1.35
−1.77*
-
-
18
Nigeria
NGA
1960−2018
59
2
1
−1.97
−1.14
1
−1.62
1
1
−3.75
3.14
1
−1.93
19
Rwanda
RWA
1960−2018
59
2
0
0.36
−1.83**
-
-
1
0
−2.90
3.68
0
−0.23
20
Sudan
SDN
1960−2018
59
1
0
1.66
−1.90**
-
-
3
0
−0.10
−1.18
0
−0.10
21
Senegal
SEN
1960−2018
59
1
1
1.03
−2.73***
-
-
3
0
0.48
1.38
0
0.44
22
Sierra Leone
SLE
1960−2018
59
2
0
−1.94
−2.05**
-
-
1
0
−4.32
4.03
0
−1.27
23
Seychelles
SYC
1960−2018
59
3
0
0.70
−2.14**
-
-
3
0
−1.83
−1.79*
-
-
24
Chad
TCD
1960−2018
59
1
0
−0.85
−2.28**
-
-
1
0
−0.77
−1.00
0
−1.48
25
Togo
TGO
1960−2018
59
1
1
−3.88
2.28
1
−2.94**
1
1
−5.62
4.17
1
−2.93
26
South Africa
ZAF
1960−2018
59
2
1
−2.34
−1.54*
1
−1.75
2
1
−2.82
−1.88**
-
-
27
Zambia
ZMB
1960−2018
59
5
2
−3.11
2.10
2
−3.06**
5
2
−1.77
1.96
2
−1.63
28
Zimbabwe
ZWE
1960−2018
59
5
2
−3.86***
−2.63**
-
-
5
2
−3.46**
−2.48**
-
-
29
Gambia, The
GMB
1966−2018
53
5
0
−2.42
1.62
0
−2.38
5
0
−2.19
1.61
3
−4.30***
30
Mali
MLI
1967−2018
52
4
0
0.12
−2.46**
-
-
4
0
−1.81
−1.90**
-
-
31
Guinea−Bissau
GNB
1970−2018
49
5
0
−2.56
−0.90
0
−2.65*
1
0
−2.47
1.05
0
−2.56
32
Eswatini
SWZ
1970−2018
49
3
0
1.19
2.42
0
0.58
3
0
−1.68
2.06
0
−2.15
33
Mauritius
MUS
1976−2018
43
4
2
3.79
4.09
3
4.09
4
2
2.10
4.37
3
0.97
34
Angola
AGO
1980−2018
39
4
1
−2.36
−2.70**
-
-
1
2
1.33
−2.74***
-
-
35
Comoros
COM
1980−2018
39
4
1
−3.02
2.01
1
−2.08
4
1
−1.87
1.87
2
−1.64
36
Cabo Verde
CPV
1980−2018
39
3
1
−0.27
2.18
1
−0.47
3
1
−2.19
1.75
1
−2.65
37
Equatorial Guinea
GNQ
1980−2018
39
1
2
−6.33***
−6.77***
-
-
4
2
−3.38**
−1.72*
-
-
38
Mozambique
MOZ
1980−2018
39
1
1
1.28
−3.43***
-
-
1
2
−0.48
−1.41*
-
-
39
Namibia
NAM
1980−2018
39
1
1
0.19
−3.42***
-
-
4
0
−2.46
−1.80**
-
-
40
Ethiopia
ETH
1981−2018
38
1
0
3.36
−3.42***
-
-
1
0
2.30
−1.77**
-
-
41
Uganda
UGA
1982−2018
37
1
1
−1.24
−2.45**
-
-
4
1
−2.57
−0.86
1
−2.58
42
Guinea
GIN
1986−2018
33
1
2
8.15
6.22
2
1.58
2
0
2.78
−3.79***
-
-
43
Tanzania
TZA
1988−2018
31
5
2
0.62
−1.29*
2
0.86
2
1
−5.04
5.04
1
−5.48***
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
In addition to the results presented in Table 2–8, the overall summary of these empirical results is presented in Table 9 to get a quick idea of the key results. These results exhibit that out of a total of 218 countries, 177 were selected for the analysis with at-least 25 years of available data and out of these 177 countries, the per capita GDP series is found to be stationary for 70, while for the rest of 107 countries it is found to be non-stationary. To get a summary containing exact names of countries for which per capita GDP is stationary, the findings are presented in the form of a bar graph (Figure 1), where a bar of unit length against a country is added, indicating the rejection of null of unit root. The empirical estimates related to stationarity of real GDP per capita income in 70 countries of this study are consistent with the findings by Narayan (2004), Murthy and Anoruo (2009), Ying et al. (2013) and Baktemur (2019), while results of non-stationarity of GDP per capita in 107 countries are in accord with Rapach (2002), Ozturk and Kalyoncu (2007), Aslanidis and Fountas (2014), Firat (2016), and Oskooe and Akbari (2015). Overall results are consistent with the findings by Smyth (2003), Chang et al.(2005, 2006), Tiwari and Suresh (2014), Emirmahmutoglu et al. (2019).
Table 9.
Summary of overall results.
Region
Total countries
Selected countries
Stationary under only "Constant" specification(A)
Stationary under only "Constant & trend" specification (B)
An important distinction of the empirical results presented in this study from the existing studies is that the tests used in present study makes use of both frequency and time dimensions into account which is ignored by all previous studies. As emphasized in Aydin and Pata (2020), ignoring information regarding frequency domain and structural change may lead to biased results. Thus, the findings presented in present study are more complete and robust and provide a better picture of the issue at hand which his ignored by previous studies considering the same issue and hence, the present study fills this much needed gap.
5.
Summary and conclusions
The role of GDP per capita cannot be overlooked measuring socio-economic status of individual and country's economic condition. Policy makers usually need stationary time series data for modeling and forecasting. They avoid using non-stationary data which are changeable and cannot be used for modeling and forecasting due to its erratic (unstable) behavior. The empirical findings derived using non-stationary time series data may be forged. The non-stationary data may be converted in to stationary data which yields more reliable, consistent results. Therefore, it is indispensable to determine whether the GDP per capita income has a trend stationary or a difference stationary process. Thus, the broad aim this study is to evaluate the stationarity properties of GDP per capita for a panel of 177 out of total 218 countries over 1960–2018 with at-least 25 years of available annual data, considering seven sub-samples (regions) categorized by world bank. We utilized historical, longer period updated data, and covered the entire world which is never evaluated by any of erstwhile studies before. We implemented the newly developed unit root tests (WADF and FWADF) by Pata and Aydin (2020) based on discrete wavelet transform to analyze the stationarity properties. An important advantage of these wavelet unit root tests is that these have the ability to consider the smooth structural break into account which is not considered by any of the previous existing wavelet tests such as the one proposed by Fan and Gencay (2010). Thus, the empirical results presented in this study are better than existing approaches and provide new insight of long-standing issue of testing the stationarity of per capita GDP series across the globe.
The empirical estimates reveal that GDP per capita is found to be stationary in 70 including 6 out of 26 East Asia and Pacific countries, 25 out of 46 Europe and Central Asian countries, 14 out of 36 Latin America and the Caribbean countries, 12 out of 17 Middle East and North African countries, 1 out of 3 three selected countries in North America, 1 out of 6 South Asian countries and 11 out of 43 countries in Sub-Saharan Africa, while for the rest of 107 countries, the real GDP per capita is found to be non-stationary. It is obvious from the empirical findings that GDP per capita income is stationary in several countries. The stationary time series of per capita GDP is desirable for any public policy across the world, where as the non-stationary GDP per capita needs to be transformed into stationary data in order to avoid inconsistency in the data. These findings suggest that it is imperative for policy makers to check the stationary properties of real GDP per capita series prior to utilizing it for any policy formulation using the empirical approaches used in this study which not only considers both freqency and time domain into account but also smooth structural changes as well.
Acknowledgements
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. The authors thank Mucahit Aydin for sharing replication codes.
Conflicts of interest
All authors declare no conflicts of interest in this paper.
References
[1]
A. S. Al-Goblan, M. A. Al-Alfi, M. Z. Khan. Mechanism linking diabetes mellitus and obesity, Diabetes Metab. Syndr. Obes., 7 (2014), 587–591. doi: 10.2147/dmso.S67400 doi: 10.2147/dmso.S67400
[2]
A. A. Rao, N. M. Tayaru, H. Thota, S. B. Changalasetty, L. S. Thota, S. Gedela, Bioinformatic analysis of functional proteins involved in obesity associated with diabetes, Int. J. Biomed. Sci., 4 (2008), 70–73.
[3]
P. E. Scherer, J. A. Hill, Obesity, diabetes, and cardiovascular diseases: A compendium, Circ. Res., 118 (2016), 1703–1705. doi: 10.1161/circresaha.116.308999 doi: 10.1161/circresaha.116.308999
[4]
G. R. Babu, G. V. S. Murthy, Y. Ana, P. Patel, R. Deepa, S. E. B. Neelon, et al. Association of obesity with hypertension and type 2 diabetes mellitus in India: A meta-analysis of observational studies, World J. Diabetes, 9 (2018), 40–52. doi: 10.4239/wjd.v9.i1.40 doi: 10.4239/wjd.v9.i1.40
[5]
A. Medina-Remón, R. Kirwan, R. M. Lamuela-Raventós, R. Estruch. Dietary patterns and the risk of obesity, type 2 diabetes mellitus, cardiovascular diseases, asthma, and neurodegenerative diseases, Crit. Rev. Food Sci. Nutr., 58 (2018), 262–296. doi: 10.1080/10408398.2016.1158690 doi: 10.1080/10408398.2016.1158690
[6]
G. A. Bray, Medical consequences of obesity, J. Clin. Endocrinol. Metab., 89 (2004), 2583–2589. doi: 10.1210/jc.2004-0535 doi: 10.1210/jc.2004-0535
[7]
J. S. M. Sabir, A. El Omri, B. Banaganapalli, N. Aljuaid, A. M. S. Omar, A. Altaf, et al., Unraveling the role of salt-sensitivity genes in obesity with integrated network biology and co-expression analysis, PLoS One, 15 (2020), e0228400. doi: 10.1371/journal.pone.0228400 doi: 10.1371/journal.pone.0228400
[8]
M. B. Zimering, V. Delic, B. A. Citron, Gene expression changes in a model neuron cell line exposed to autoantibodies from patients with traumatic brain injury and/or Type 2 diabetes, Mol. Neurobiol., (2021). doi: 10.1007/s12035-021-02428-4 doi: 10.1007/s12035-021-02428-4
[9]
T. O. Kilpeläinen, T. A. Lakka, D. E. Laaksonen, J. Lindström, J. G. Eriksson, T. T. Valle, et al., SNPs in PPARG associate with type 2 diabetes and interact with physical activity, Med. Sci. Sports Exerc., 40 (2008), 25–33. doi: 10.1249/mss.0b013e318159d1cd doi: 10.1249/mss.0b013e318159d1cd
[10]
J. J. Jia, X. Zhang, C. R. Ge, M. Jois, The polymorphisms of UCP2 and UCP3 genes associated with fat metabolism, obesity and diabetes, Obes. Rev., 10 (2009), 519–526. doi: 10.1111/j.1467-789X.2009.00569.x doi: 10.1111/j.1467-789X.2009.00569.x
[11]
D. Meyre, N. Bouatia-Naji, A. Tounian, C. Samson, C. Lecoeur, V. Vatin, et al., Variants of ENPP1 are associated with childhood and adult obesity and increase the risk of glucose intolerance and type 2 diabetes, Nat. Genet., 37 (2005), 863–867. doi: 10.1038/ng1604 doi: 10.1038/ng1604
[12]
T. M. Frayling, N. J. Timpson, M. N. Weedon, E. Zeggini, R. M. Freathy, C. M. Lindgren, et al., A common variant in the FTO gene is associated with body mass index and predisposes to childhood and adult obesity, Science, 316 (2007), 889–894. doi: 10.1126/science.1141634 doi: 10.1126/science.1141634
[13]
M. Hong, S. Tao, L. Zhang, L.-T. Diao, X. Huang, S. Huang, et al., RNA sequencing: New technologies and applications in cancer research, J. Hematol. Oncol., 13 (2020), 166. doi: 10.1186/s13045-020-01005-x doi: 10.1186/s13045-020-01005-x
[14]
G. Laenen, L. Thorrez, D. Börnigen, Y. Moreau, Finding the targets of a drug by integration of gene expression data with a protein interaction network, Mol. Biosyst., 9 (2013), 1676–1685. doi: 10.1039/c3mb25438k doi: 10.1039/c3mb25438k
[15]
R. Roy, L. N. Winteringham, T. Lassmann, A. R. R. Forrest. Expression levels of therapeutic targets as indicators of sensitivity to targeted therapeutics, Mol. Cancer Ther., 18 (2019), 2480–2489. doi: 10.1158/1535-7163.Mct-19-0273 doi: 10.1158/1535-7163.Mct-19-0273
[16]
R. Edgar, M. Domrachev, A. E. Lash, Gene expression omnibus: NCBI gene expression and hybridization array data repository, Nucleic Acids Res., 30 (2002), 207–210. doi: 10.1093/nar/30.1.207 doi: 10.1093/nar/30.1.207
[17]
S. W. Wingett, S. Andrews, FastQ screen: A tool for multi-genome mapping and quality control, F1000Res, 7 (2018), 1338. doi: 10.12688/f1000research.15931.2 doi: 10.12688/f1000research.15931.2
[18]
A. M. Bolger, M. Lohse, B. Usadel, Trimmomatic: A flexible trimmer for Illumina sequence data, Bioinformatics, 30 (2014), 2114–2120. doi: 10.1093/bioinformatics/btu170 doi: 10.1093/bioinformatics/btu170
[19]
A. Dobin, C. A. Davis, F. Schlesinger, J. Drenkow, C. Zaleski, S. Jha, et al., STAR: Ultrafast universal RNA-seq aligner, Bioinformatics, 29 (2013), 15–21. doi: 10.1093/bioinformatics/bts635 doi: 10.1093/bioinformatics/bts635
[20]
Y. Liao, G. K. Smyth, W. Shi, featureCounts: An efficient general purpose program for assigning sequence reads to genomic features, Bioinformatics, 30 (2014), 923–930. doi: 10.1093/bioinformatics/btt656 doi: 10.1093/bioinformatics/btt656
[21]
M. I. Love, W. Huber, S. Anders, Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2, Genome Biol., 15 (2014), 550. doi: 10.1186/s13059-014-0550-8 doi: 10.1186/s13059-014-0550-8
[22]
R. Kolde, pheatmap: Pretty Heatmaps. R package version 1.0. 8, in, Available, 2015.
[23]
K. Blighe, S. Rana, M. Lewis, EnhancedVolcano: Publication-ready volcano plots with enhanced colouring and labeling (2019), R Package Version, 1 (2018).
[24]
M. H. Schaefer, J. F. Fontaine, A. Vinayagam, P. Porras, E. E. Wanker, M. A. Andrade-Navarro, HIPPIE: Integrating protein interaction networks with experiment based quality scores, PLoS One, 7 (2012), e31826. doi: 10.1371/journal.pone.0031826 doi: 10.1371/journal.pone.0031826
[25]
G. Alanis-Lobato, M. A. Andrade-Navarro, M. H. Schaefer, HIPPIE v2.0: Enhancing meaningfulness and reliability of protein-protein interaction networks, Nucleic Acids Res., 45 (2017), D408–D414. doi: 10.1093/nar/gkw985 doi: 10.1093/nar/gkw985
[26]
P. Shannon, A. Markiel, O. Ozier, N. S. Baliga, J. T. Wang, D. Ramage, et al., Cytoscape: A software environment for integrated models of biomolecular interaction networks, Genome Res., 13 (2003), 2498–2504. doi: 10.1101/gr.1239303 doi: 10.1101/gr.1239303
[27]
Y. Tang, M. Li, J. Wang, Y. Pan, F. X. Wu. CytoNCA: A cytoscape plugin for centrality analysis and evaluation of protein interaction networks, Biosystems, 127 (2015), 67–72. doi: 10.1016/j.biosystems.2014.11.005 doi: 10.1016/j.biosystems.2014.11.005
[28]
S. Wasserman, K. Faust, Social network analysis: Methods and applications, (1994).
[29]
S. P. Borgatti, Centrality and network flow, Social networks, 27 (2005), 55–71. doi: 10.1016/j.socnet.2004.11.008 doi: 10.1016/j.socnet.2004.11.008
[30]
L. C. Freeman, Centrality in social networks conceptual clarification, Social networks, 1 (1978), 215–239. doi: 10.1016/0378-8733(78)90021-7 doi: 10.1016/0378-8733(78)90021-7
[31]
M. E. Newman, The mathematics of networks, The new palgrave encyclopedia of economics, 2 (2008), 1–12.
[32]
G. George, S. Valiya Parambath, S. B. Lokappa, J. Varkey, Construction of Parkinson's disease marker-based weighted protein-protein interaction network for prioritization of co-expressed genes, Gene, 697 (2019), 67–77. doi: 10.1016/j.gene.2019.02.026 doi: 10.1016/j.gene.2019.02.026
[33]
C. Durón, Y. Pan, D. H. Gutmann, J. Hardin, A. Radunskaya, Variability of betweenness centrality and its effect on identifying essential genes, Bull. Math. Biol., 81 (2019), 3655–3673. doi: 10.1007/s11538-018-0526-z doi: 10.1007/s11538-018-0526-z
[34]
J. Chen, E. E. Bardes, B. J. Aronow, A. G. Jegga, ToppGene Suite for gene list enrichment analysis and candidate gene prioritization, Nucleic Acids Res., 37 (2009), W305–311. doi: 10.1093/nar/gkp427 doi: 10.1093/nar/gkp427
[35]
C. S. Greene, A. Krishnan, A. K. Wong, E. Ricciotti, R. A. Zelaya, D. S. Himmelstein, et al., Understanding multicellular function and disease with human tissue-specific networks, Nat. Genet., 47 (2015), 569–576. doi: 10.1038/ng.3259 doi: 10.1038/ng.3259
[36]
G. Koscielny, P. An, D. Carvalho-Silva, J. A. Cham, L. Fumis, R. Gasparyan, et al., Open Targets: A platform for therapeutic target identification and validation, Nucleic Acids Res., 45 (2017), D985–d994. doi: 10.1093/nar/gkw1055 doi: 10.1093/nar/gkw1055
[37]
C. L. Haase, A. Tybjærg-Hansen, B. G. Nordestgaard, R. Frikke-Schmidt, HDL cholesterol and risk of Type 2 diabetes: A mendelian randomization study, Diabetes, 64 (2015), 3328–3333. doi: 10.2337/db14-1603 doi: 10.2337/db14-1603
[38]
M. A. Javed Shaikh, R. S. H. Singh, S. Rawat, S. Pathak, A. Mishra, et al., Role of various gene expressions in etiopathogenesis of Type 2 diabetes mellitus, Adv. Mind. Body Med., 35 (2021), 31 –39. PMID: 34237027.
[39]
T. Liu, J. Liu, L. Hao, Network pharmacological study and molecular docking analysis of qiweitangping in treating diabetic coronary heart disease, Evid. Based Complement. Alternat. Med., 2021 (2021), 9925556. doi: 10.1155/2021/9925556 doi: 10.1155/2021/9925556
[40]
N. N. Sahly, B. Banaganapalli, A. N. Sahly, A. H. Aligiraigri, K. K. Nasser, T. Shinawi, et al., Molecular differential analysis of uterine leiomyomas and leiomyosarcomas through weighted gene network and pathway tracing approaches, Syst. Biol. Reprod. Med., 67 (2021), 209–220. doi: 10.1080/19396368.2021.1876179 doi: 10.1080/19396368.2021.1876179
[41]
B. Banaganapalli, N. Al-Rayes, Z. A. Awan, F. A. Alsulaimany, A. S. Alamri, R. Elango, et al., Multilevel systems biology analysis of lung transcriptomics data identifies key miRNAs and potential miRNA target genes for SARS-CoV-2 infection, Comput. Biol. Med., 135 (2021), 104570. doi: 10.1016/j.compbiomed.2021.104570 doi: 10.1016/j.compbiomed.2021.104570
[42]
A. Mujalli, B. Banaganapalli, N. M. Alrayes, N. A. Shaik, R. Elango, J. Y. Al-Aama, Myocardial infarction biomarker discovery with integrated gene expression, pathways and biological networks analysis, Genomics, 112 (2020), 5072–5085. doi: 10.1016/j.ygeno.2020.09.004 doi: 10.1016/j.ygeno.2020.09.004
[43]
T. Ideker, R. Nussinov, Network approaches and applications in biology, PLoS Comput. Biol., 13 (2017), e1005771–e1005771. doi: 10.1371/journal.pcbi.1005771 doi: 10.1371/journal.pcbi.1005771
[44]
D. O. Holland, B. H. Shapiro, P. Xue, M. E. Johnson, Protein-protein binding selectivity and network topology constrain global and local properties of interface binding networks, Sci. Rep., 7 (2017), 5631. doi: 10.1038/s41598-017-05686-2 doi: 10.1038/s41598-017-05686-2
[45]
Y. Gao, X. Chang, J. Xia, S. Sun, Z. Mu, X. Liu, Identification of HCC-related genes based on differential partial correlation network, Front Genet, 12 (2021), 672117. doi: 10.3389/fgene.2021.672117 doi: 10.3389/fgene.2021.672117
[46]
C. Liu, L. Lu, Q. Kong, Y. Li, H. Wu, W. Yang, et al., Developing discriminate model and comparative analysis of differentially expressed genes and pathways for bloodstream samples of diabetes mellitus type 2, BMC Bioinform., 15 Suppl 17 (2014), S5. doi: 10.1186/1471-2105-15-s17-s5 doi: 10.1186/1471-2105-15-s17-s5
[47]
G. Prashanth, B. Vastrad, A. Tengli, C. Vastrad, I. Kotturshetti, Investigation of candidate genes and mechanisms underlying obesity associated type 2 diabetes mellitus using bioinformatics analysis and screening of small drug molecules, BMC Endocr. Disord., 21 (2021), 80. doi: 10.1186/s12902-021-00718-5 doi: 10.1186/s12902-021-00718-5
[48]
X. Yao, J. Yan, K. Liu, S. Kim, K. Nho, S. L. Risacher, et al., Tissue-specific network-based genome wide study of amygdala imaging phenotypes to identify functional interaction modules, Bioinformatics, 33 (2017), 3250–3257. doi: 10.1093/bioinformatics/btx344 doi: 10.1093/bioinformatics/btx344
[49]
R. L. J. van Meijel, E. E. Blaak, G. H. Goossens, Chapter 1 - Adipose tissue metabolism and inflammation in obesity, in: R. A. Johnston, B. T. Suratt (Eds.), Mechanisms and Manifestations of Obesity in Lung Disease, Academic Press, 2019, pp. 1–22.
[50]
C. Fotis, A. Antoranz, D. Hatziavramidis, T. Sakellaropoulos, L. G. Alexopoulos, Network-based technologies for early drug discovery, Drug Discovery Today, 23 (2018), 626–635. doi: 10.1016/j.drudis.2017.12.001 doi: 10.1016/j.drudis.2017.12.001
[51]
J. M. Fernandez-Real, J. A. Menendez, J. M. Moreno-Navarrete, M. Blüher, A. Vazquez-Martin, M. J. Vázquez, et al., Extracellular fatty acid synthase: A possible surrogate biomarker of insulin resistance, Diabetes, 59 (2010), 1506–1511. doi: 10.2337/db09-1756 doi: 10.2337/db09-1756
[52]
A. Ray, Tumor-linked HER2 expression: Association with obesity and lipid-related microenvironment, Horm. Mol. Biol. Clin. Investig., 32 (2017). doi: 10.1515/hmbci-2017-0020 doi: 10.1515/hmbci-2017-0020
[53]
F. J. Ruiz-Ojeda, A. Méndez-Gutiérrez, C. M. Aguilera, J. Plaza-Díaz, Extracellular matrix remodeling of adipose tissue in obesity and metabolic diseases, Int. J. Mol. Sci., 20 (2019). doi: 10.3390/ijms20194888 doi: 10.3390/ijms20194888
[54]
P. Järgen, A. Dietrich, A. W. Herling, H. P. Hammes, P. Wohlfar, The role of insulin resistance in experimental diabetic retinopathy-Genetic and molecular aspects, PLoS One, 12 (2017), e0178658. doi: 10.1371/journal.pone.0178658 doi: 10.1371/journal.pone.0178658
[55]
M. C. Tse, X. Liu, S. Yang, K. Ye, C. B. Chan, Fyn regulates adipogenesis by promoting PIKE-A/STAT5a interaction, Mol. Cell. Biol., 33 (2013), 1797–1808. doi: 10.1128/mcb.01410-12 doi: 10.1128/mcb.01410-12
[56]
C. C. Bastie, H. Zong, J. Xu, B. Busa, S. Judex, I. J. Kurland, et al., Integrative metabolic regulation of peripheral tissue fatty acid oxidation by the SRC kinase family member Fyn, Cell Metab., 5 (2007), 371–381. doi: 10.1016/j.cmet.2007.04.005 doi: 10.1016/j.cmet.2007.04.005
[57]
E. Yamada, J. E. Pessin, I. J. Kurland, G. J. Schwartz, C. C. Bastie, Fyn-dependent regulation of energy expenditure and body weight is mediated by tyrosine phosphorylation of LKB1, Cell Metab., 11 (2010), 113–124. doi: 10.1016/j.cmet.2009.12.010 doi: 10.1016/j.cmet.2009.12.010
[58]
C. C. Bastie, H. H. Zong, J. Xu, S. Judex, I. J. Kurland, J. E. Pessin, Fyn kinase deficiency increases peripheral tissue insulin sensitivity by improving fatty acid oxidation and lipolysis, Diabetes, 56 (2007), A60.
[59]
J. Rodrigues-Krause, M. Krause, C. O'Hagan, G. De Vito, C. Boreham, C. Murphy, et al., Divergence of intracellular and extracellular HSP72 in type 2 diabetes: Does fat matter?, Cell Stress Chaperones, 17 (2012), 293–302. doi: 10.1007/s12192-011-0319-x doi: 10.1007/s12192-011-0319-x
[60]
P. L. Hooper, G. Balogh, E. Rivas, K. Kavanagh, L. Vigh, The importance of the cellular stress response in the pathogenesis and treatment of type 2 diabetes, Cell Stress Chaperones, 19 (2014), 447–464. doi: 10.1007/s12192-014-0493-8 doi: 10.1007/s12192-014-0493-8
[61]
E. Chang, M. Varghese, K. Singer, Gender and sex differences in adipose tissue, Curr. Diab. Rep., 18 (2018), 69. doi: 10.1007/s11892-018-1031-3 doi: 10.1007/s11892-018-1031-3
This article has been cited by:
1.
Yue Liu, Liming Chen, Liangting Lv, Pierre Failler,
The impact of population aging on economic growth: a case study on China,
2023,
8,
2473-6988,
10468,
10.3934/math.2023531
2.
Mumtaz Ahmed, Muhammad Azam, Atif Maqbool Khan,
Testing stationarity of per capita energy consumption across the globe: a better picture using the wavelet framework,
2022,
29,
0944-1344,
65334,
10.1007/s11356-022-20416-6
3.
Liu Hong, Li Lin, Xiaohang Ren,
Evaluation of Interest Balance of Low-carbon Collaborative Innovation Subjects,
2022,
2022,
1563-5147,
1,
10.1155/2022/8270712
4.
Hao Dong, Zhehao Huang,
Decomposing and reconstructing dynamic risks in the crude oil market based on the VMD and Lempel–Ziv algorithms,
2022,
30,
2688-1594,
4674,
10.3934/era.2022237
5.
Mumtaz Ahmed, Zulekha Qadeer, Muhammad Azam Khan, Toqeer Ahmed,
Evaluating external debt sustainability across the globe: evidence from wavelet-based tests,
2025,
0033-5177,
10.1007/s11135-025-02053-6
Abdulhadi Ibrahim H. Bima, Ayman Zaky Elsamanoudy, Walaa F Albaqami, Zeenath Khan, Snijesh Valiya Parambath, Nuha Al-Rayes, Prabhakar Rao Kaipa, Ramu Elango, Babajan Banaganapalli, Noor A. Shaik. Integrative system biology and mathematical modeling of genetic networks identifies shared biomarkers for obesity and diabetes[J]. Mathematical Biosciences and Engineering, 2022, 19(3): 2310-2329. doi: 10.3934/mbe.2022107
Abdulhadi Ibrahim H. Bima, Ayman Zaky Elsamanoudy, Walaa F Albaqami, Zeenath Khan, Snijesh Valiya Parambath, Nuha Al-Rayes, Prabhakar Rao Kaipa, Ramu Elango, Babajan Banaganapalli, Noor A. Shaik. Integrative system biology and mathematical modeling of genetic networks identifies shared biomarkers for obesity and diabetes[J]. Mathematical Biosciences and Engineering, 2022, 19(3): 2310-2329. doi: 10.3934/mbe.2022107
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
S. No.
Country
Code
Time span
sample size
With Constant
With Constant and Trend
k
p
FWKSS
t-stat
p
WADF
k
p
FWKSS
t-stat
p
WADF
1
Austria
AUT
1960–2018
59
1
0
−2.39
−1.79*
-
-
1
1
−1.76
−1.89**
-
-
2
Belgium
BEL
1960–2018
59
2
0
−0.81
1.56
0
−1.55
1
1
−2.38
−1.36
1
−2.18
3
Denmark
DNK
1960–2018
59
2
0
−0.60
1.75
0
−1.31
1
0
−3.32
−2.42**
-
-
4
Spain
ESP
1960–2018
59
2
2
−0.58
2.16
2
−1.11
3
1
−4.10
0.89
1
-4.16***
5
Finland
FIN
1960–2018
59
3
0
−0.25
2.52
0
−0.72
3
1
−3.12
1.71
1
-3.47**
6
France
FRA
1960–2018
59
2
0
−2.13
1.51
0
−2.94**
2
0
0.02
1.50
0
−0.52
7
United Kingdom
GBR
1960–2018
59
1
0
−1.53
−1.96**
-
-
1
0
−2.56
−2.28**
-
-
8
Greece
GRC
1960–2018
59
1
1
−3.05*
−1.95*
-
-
3
1
−3.09
1.28
1
-3.17*
9
Iceland
ISL
1960–2018
59
3
0
0.21
1.66
0
−0.12
2
1
−4.70***
−2.68***
-
-
10
Italy
ITA
1960–2018
59
2
1
−1.70
1.24
1
−1.94
2
0
1.04
1.37
0
1.06
11
Luxembourg
LUX
1960–2018
59
1
0
−2.21
−2.89***
-
-
1
1
−3.08
−3.25***
-
-
12
Netherlands
NLD
1960–2018
59
1
1
−2.16
−2.18**
-
-
5
1
−2.67
1.10
1
−2.70
13
Norway
NOR
1960–2018
59
1
1
−1.92
−1.54*
1
−1.17
1
1
−3.47*
−3.18***
-
-
14
Portugal
PRT
1960–2018
59
2
2
−1.53
2.85
2
−1.72
2
1
0.34
2.78
1
-2.57
15
Sweden
SWE
1960–2018
59
3
0
0.81
2.55
0
0.17
3
0
−1.20
2.30
0
-1.62
16
Turkey
TUR
1960–2018
59
3
0
2.97
−0.67
0
3.30
3
0
2.08
−1.56*
-
-
17
Georgia
GEO
1965–2018
54
2
1
−3.03
−1.52*
1
−2.65*
1
1
−4.38
3.13
1
−2.57
18
Andorra
AND
1970–2018
49
1
1
−1.66
−1.85**
-
-
2
1
−1.15
1.95
1
−3.09*
19
Germany
DEU
1970–2018
49
4
0
−0.56
−1.39
0
−0.42
4
0
−3.11*
−1.58*
-
-
20
Greenland
GRL
1970–2018
49
1
0
−0.75
−1.28
0
0.01
1
1
−2.19
1.41
1
−1.86
21
Ireland
IRL
1970–2018
49
4
1
0.71
−1.38
3
1.33
2
1
−2.67
1.70
1
−2.49
22
Monaco
MCO
1970–2018
49
4
2
2.44
−1.84**
-
-
4
2
0.59
−1.80**
-
-
23
Cyprus
CYP
1975–2018
44
2
0
−1.69
1.97
0
−1.56
1
1
−2.35
−1.41*
-
-
24
Albania
ALB
1980–2018
39
1
1
−0.49
−4.54***
-
-
4
1
−2.21
−1.21
3
−3.79**
25
Bulgaria
BGR
1980–2018
39
2
0
3.12
3.55
0
1.23
4
1
−1.61
−2.49**
-
-
26
Isle of Man
IMN
1984–2017
34
4
0
1.67
−2.51**
-
-
2
2
−2.25
1.95
2
-2.61
27
Tajikistan
TJK
1985–2018
34
2
1
−3.35**
−2.14**
-
-
2
2
−4.90
0.97
3
−10.35***
28
Kyrgyz Republic
KGZ
1986–2018
33
1
1
−0.37
−3.23***
-
-
4
2
−5.67
−0.91
3
−9.98***
29
Turkmenistan
TKM
1987–2018
32
1
1
3.00
−4.31***
-
-
5
2
−2.96
2.84
2
−1.91
30
Ukraine
UKR
1887–2018
32
1
1
−3.89***
−2.29**
-
-
5
2
−5.31
1.67
3
−5.08***
31
Uzbekistan
UZB
1987–2018
32
1
2
4.39
−6.00***
-
-
1
2
0.51
−3.48***
-
-
32
Russian Federation
RUS
1989–2018
30
5
2
−2.09
1.53
3
-2.97**
5
2
−3.25
2.16
3
−2.92
33
Armenia
ARM
1990–2018
29
3
1
−2.14
−3.07***
-
-
3
2
−1.51
−1.99**
-
-
34
Azerbaijan
AZE
1990–2018
29
1
1
−3.24
−4.00***
-
-
1
1
−4.60***
−4.27***
-
-
35
Belarus
BLR
1990–2018
29
1
2
−10.15***
−9.70***
-
-
1
1
−2.25
−10.7***
-
-
36
Czech Republic
CZE
1990–2018
29
2
1
−0.52
3.95
1
−0.49
2
2
−3.89
3.81
3
−3.97**
37
Kazakhstan
KAZ
1990–2018
29
5
2
−2.38
1.82
2
−2.22
1
1
−5.00**
−3.00***
-
-
38
North Macedonia
MKD
1990–2018
29
1
1
0.35
−1.96**
-
-
1
0
−2.14
−1.31*
0
−5.57***
39
Romania
ROU
1990–2018
29
2
0
3.32
5.05
1
0.48
2
2
0.19
2.63
3
−3.17*
40
Slovenia
SVN
1990–2018
29
2
1
−2.09
1.73
3
−1.74
1
1
−6.36***
−4.68***
-
-
41
Hungary
HUN
1991–2018
28
2
1
−0.25
−1.57**
1
−0.39
2
1
−4.46
2.48
1
−3.52**
42
Slovak Republic
SVK
1992–2018
27
2
0
0.88
4.94
0
0.20
3
2
−5.01***
−2.62**
-
-
43
Estonia
EST
1993–2018
26
4
0
−0.83
1.45
0
−0.76
2
0
−3.39
2.59
0
−1.97
44
Bosnia and Herzegovina
BIH
1994–2018
25
1
2
−2.29
3.30
2
−2.28
2
2
−5.44
2.75
3
−5.91***
45
Latvia
LVA
1994–2018
25
2
0
−1.76
1.87
0
−0.56
2
0
−4.69
4.77
1
−2.33
46
Serbia
SRB
1994–2018
25
5
2
−1.47
−1.52**
2
−1.66
4
2
−4.76
2.23
2
−3.34*
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
S. No.
Country
Code
Time span
sample size
With Constant
With Constant and Trend
k
p
FWKSS
t-stat
P
WADF
k
p
FWKSS
t-stat
P
WADF
1
Argentina
ARG
1960–2018
59
4
0
−1.12
1.07
0
−0.67
1
0
−3.44
2.33
1
−2.55
2
Bahamas
BHS
1960–2018
59
1
0
−4.46***
−3.41***
-
-
1
1
−4.26**
−2.64***
-
-
3
Belize
BLZ
1960–2018
59
1
1
−2.90
−2.83***
-
-
1
1
−4.21**
−3.05***
-
-
4
Bolivia
BOL
1960–2018
59
3
1
−0.65
−2.22***
-
-
1
1
−5.32
4.61
1
−2.47
5
Brazil
BRA
1960–2018
59
5
1
−1.73
1.78
1
−1.72
2
1
−3.97**
−2.81***
-
-
6
Chile
CHL
1960–2018
59
1
0
−0.21
−3.28***
-
-
1
0
−0.29
−2.29**
-
-
7
Colombia
COL
1960–2018
59
3
2
1.95
−2.64***
-
-
1
1
−3.70
2.61
1
−2.37
8
Costa Rica
CRI
1960–2018
59
3
0
1.89
−2.02**
-
-
3
0
−0.10
−2.03***
-
-
9
Dominican Republic
DOM
1960–2018
59
5
0
4.34
1.86
0
3.99
5
0
1.60
1.87
0
1.31
10
Ecuador
ECU
1960–2018
59
2
1
−1.30
−2.23**
-
-
2
1
−4.08**
−2.63***
-
-
11
Guatemala
GTM
1960–2018
59
3
1
−1.14
−2.29**
-
-
1
1
−3.82
2.15
1
−2.95*
12
Guyana
GUY
1960–2018
59
3
1
2.15
−3.28***
-
-
3
1
0.35
−3.06***
-
-
13
Honduras
HND
1960–2018
59
4
0
0.33
1.40
0
0.10
1
0
−5.98
5.41
0
−1.32
14
Haiti
HTI
1960–2018
59
1
1
−2.96
2.71
1
−1.35
2
1
−3.20
−3.66***
-
-
15
Mexico
MEX
1960–2018
59
5
0
−1.80
−1.90**
-
-
5
0
−1.83
−1.68*
-
-
16
Nicaragua
NIC
1960–2018
59
1
1
−2.02
1.46
1
−1.68
1
1
−2.51
2.04
1
−1.45
17
Panama
PAN
1960–2018
59
3
0
3.15
−0.62
0
3.45
1
0
−1.17
1.48
0
1.39
18
Peru
PER
1960–2018
59
1
1
−0.30
−2.32**
-
-
4
1
−1.60
−1.03
1
−0.43
19
Puerto Rico
PRI
1960–2018
59
1
0
−2.57
−2.27**
-
-
1
0
−1.75
−2.31**
-
-
20
Paraguay
PRY
1960–2018
59
4
1
−0.08
1.88
1
−0.31
4
1
−1.94
1.33
1
−2.44
21
Suriname
SUR
1960–2018
59
2
1
−4.04***
−2.58***
-
-
2
1
−5.17***
−3.62***
-
-
22
Trinidad and Tobago
TTO
1960–2018
59
1
1
−2.79
−2.26**
-
-
5
1
−3.15*
−2.07**
-
-
23
Uruguay
URY
1960–2018
59
3
0
1.55
−3.06***
-
-
3
1
−0.68
−2.96***
-
-
24
St. Vincent and the Grenadines
VCT
1960–2018
59
1
0
−1.37
−2.63***
-
-
1
0
−2.69
−2.06**
-
-
25
Venezuela, RB
VEN
1960–2018
59
1
1
−3.21
1.94
1
−2.44*
2
1
−3.85**
−2.06**
-
-
26
El Salvador
SLV
1965–2018
54
4
1
−1.05
−1.45
1
−1.10
4
1
−1.71
−1.42
1
−1.77
27
Jamaica
JAM
1966–2018
53
1
1
−3.72**
−2.43***
-
-
2
1
−4.14
2.25
3
−4.81***
28
Switzerland
CHE
1970–2018
49
3
0
0.36
2.19
0
−0.16
2
1
−4.87**
−1.91**
-
-
29
Cuba
CUB
1970–2018
49
4
1
−0.78
1.23
1
−0.95
1
1
−3.20
2.37
1
−2.04
30
Barbados
BRB
1974–2018
45
1
0
−2.30*
−1.85**
-
-
5
1
−2.21
−2.32**
-
-
31
Antigua and Barbuda
ATG
1997–2018
42
5
0
−1.86
−1.43
0
−1.85
5
0
−1.80
−1.35*
0
−1.87
32
Dominica
DMA
1977–2018
42
1
1
−3.67**
−2.40**
-
-
1
1
−1.09
−2.35**
-
-
33
Grenada
GRD
1977–2018
42
2
0
0.18
1.60
2
−0.16
2
1
−4.31
1.15
1
−4.35**
34
St. Kitts and Nevis
KNA
1977–2018
42
3
0
−1.80
−0.99
0
−1.73
3
0
−0.21
−0.93
0
−0.62
35
St. Lucia
LCA
1977–2018
42
2
0
−2.57
−1.74*
-
-
3
2
−2.68
1.00
3
−2.52
36
Aruba
ABW
1986–2017
32
5
1
−1.66
1.60
0
−6.00***
4
0
−7.33***
−2.30***
-
-
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
S. No.
Country
Code
Time span
sample size
With Constant
With Constant and Trend
k
P
FWKSS
t-stat
P
WADF
k
p
FWKSS
t-stat
p
WADF
1
Algeria
DZA
1960–2018
59
1
2
−2.09
−1.31
2
−1.60
2
2
−6.60***
−6.31***
-
-
2
Egypt, Arab Rep.
EGY
1960–2018
59
2
2
1.80
−2.47***
-
-
2
2
−2.69*
−2.39**
-
-
3
Israel
ISR
1960–2018
59
3
0
−0.83
−2.34**
-
-
3
0
−3.20*
−2.99***
-
-
4
Iran, Islamic Rep.
IRN
1960–2017
58
1
1
−2.76
1.10
1
-2.52*
1
1
−3.07
1.72
1
−2.49
5
Oman
OMN
1965–2018
54
1
0
−4.39***
−2.57***
-
-
1
0
−2.80
−2.49**
-
-
6
Tunisia
TUN
1965–2018
54
2
0
0.58
1.54
0
0.81
2
1
−0.73
0.81
1
−1.55
7
Morocco
MAR
1966–2018
53
1
0
0.27
−1.25
1
1.33
2
0
−1.51
−1.15
1
−1.36
8
Iraq
IRQ
1968–2018
51
5
0
−0.27
1.88
0
−0.35
1
0
−3.60
2.44
0
−2.14
9
Saudi Arabia
SAU
1986–2018
51
4
2
−3.12**
−1.94**
-
-
4
2
−2.85
−1.79**
-
-
10
Malta
MLT
1970–2018
49
2
2
1.23
2.13
2
0.59
5
1
−3.74**
−1.83**
-
-
11
United Arab Emirates
ARE
1975–2018
44
2
2
−3.21
1.10
2
-3.00**
2
2
−3.90
1.84
2
−3.23*
12
Jordan
JOR
1975–2018
44
1
1
−3.14*
−1.81**
-
-
4
1
−2.69
−0.76
1
−3.13*
13
Bahrain
BHR
1980–2018
39
5
2
−3.86***
−2.63**
-
-
5
2
−3.46**
−2.48**
-
-
14
Lebanon
LBN
1988–2018
31
1
1
−3.92**
−2.17**
-
-
3
1
−3.77
2.99
3
−3.38*
15
Yemen, Rep.
YEM
1990–2018
29
4
1
−1.70
−1.36*
-
-
4
2
2.33
−1.87**
-
-
16
Kuwait
KWT
1992–2018
27
3
1
−2.13
2.02
1
−2.05
1
1
−3.52
−2.95***
-
-
17
West Bank and Gaza
PSE
1994–2018
25
1
0
1.01
−3.93***
-
-
3
2
−1.95
1.11
3
−3.26*
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
S. No.
Country
Code
Time span
samplesize
With Constant
With Constant and Trend
k
P
FWKSS
t-stat
p
WADF-stat
k
p
FWKSS
t-stat
p
WADF
1
Bermuda
BMU
1960–2013
54
1
0
−2.90
−2.07**
-
-
2
1
−4.22*
−3.19***
-
-
2
Canada
CAN
1960–2018
59
3
0
0.24
1.69
0
−0.18
2
0
−2.63
−1.98**
-
-
3
United States
USA
1960–2018
59
3
0
0.13
1.57
0
−0.18
3
0
−2.44
2.00
0
−2.14
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
S. No.
Country
Code
Time span
sample size
With Constant
With Constant and Trend
K
p
FWKSS
t-stat
p
WADF
k
p
FWKSS
t-stat
p
WADF
1
Bangladesh
BGD
1960–2018
59
4
0
9.11
1.46
0
8.82
4
0
4.10
1.38
0
3.85
2
Bhutan
BTN
1980–2018
39
1
0
3.48
−1.55*
0
5.65
5
0
0.63
1.31
0
0.45
3
India
IND
1960–2018
59
5
2
6.31
1.35
2
6.08
5
2
5.81
1.30
2
5.58
4
Sri Lanka
LKA
1961–2018
58
2
0
6.56
−1.49*
0
7.41
2
0
1.50
−1.35
0
1.53
5
Nepal
NPL
1960–2018
59
3
0
7.58
−1.20
0
7.97
3
0
2.08
−1.17
0
2.14
6
Pakistan
PAK
1960–2018
59
5
1
0.43
−1.62*
-
-
2
1
−3.93
1.84
1
−3.37**
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
S. No.
Country
Code
Time span
sample size
With Constant
With Constant and Trend
k
p
FWKSS
t-stat
p
WADF
k
p
FWKSS
t-stat
p
WADF
1
Burundi
BDI
1960–2018
59
1
0
−1.46
2.68
0
−1.41
2
0
−1.53
−2.71***
-
-
2
Benin
BEN
1960−2018
59
4
0
1.95
−2.21**
-
−
4
0
−1.37
−1.68*
-
-
3
Burkina Faso
BFA
1960−2018
59
1
0
2.20
−1.24
0
3.86
3
0
0.35
1.09
0
0.30
4
Botswana
BWA
1960−2018
59
2
0
1.00
−1.70*
-
−
3
0
−3.08
−1.12
0
−3.11*
5
Central African Republic
CAF
1960−2018
59
1
0
−2.57
2.41
0
−0.68
2
1
−5.37***
−3.81***
-
-
6
Cote d'Ivoire
CIV
1960−2018
59
1
1
−4.25
3.65
1
−1.84
1
1
−4.15
3.03
1
−2.56
7
Cameroon
CMR
1960−2018
59
5
2
−2.93
1.26
1
−2.16
5
2
−3.30
1.32
2
−3.20*
8
Congo, Dem. Rep.
COD
1960−2018
59
4
1
−1.51
−3.29***
-
-
1
2
−5.05
4.08
2
−2.27
9
Congo, Rep.
COG
1960−2018
59
2
1
−2.33
−2.82**
-
-
2
1
−0.67
−3.01***
-
-
10
Gabon
GAB
1960−2018
59
3
0
−2.37
−1.69*
-
-
3
0
−2.00
−1.91**
-
-
11
Ghana
GHA
1960−2018
59
1
0
2.18
−3.38***
-
-
3
0
0.39
−1.25
0
0.61
12
Kenya
KEN
1960−2018
59
3
0
−1.12
−1.54*
0
−0.68
1
0
−3.92
4.73
0
−1.35
13
Lesotho
LSO
1960−2018
59
4
0
3.26
2.45
0
2.60
1
0
−1.45
1.02
0
−0.10
14
Madagascar
MDG
1960−2018
59
2
0
−1.49
1.48
0
−1.21
2
0
−0.01
1.50
0
−0.76
15
Mauritania
MRT
1960−2018
59
2
2
−1.32
−0.73
2
−1.12
1
2
−1.86
1.49
3
−1.93
16
Malawi
MWI
1960−2018
59
3
0
−0.98
−2.12**
-
-
1
0
−4.17
3.37
0
−1.37
17
Niger
NER
1960−2018
59
5
2
−3.24**
−2.33**
-
-
5
1
−1.35
−1.77*
-
-
18
Nigeria
NGA
1960−2018
59
2
1
−1.97
−1.14
1
−1.62
1
1
−3.75
3.14
1
−1.93
19
Rwanda
RWA
1960−2018
59
2
0
0.36
−1.83**
-
-
1
0
−2.90
3.68
0
−0.23
20
Sudan
SDN
1960−2018
59
1
0
1.66
−1.90**
-
-
3
0
−0.10
−1.18
0
−0.10
21
Senegal
SEN
1960−2018
59
1
1
1.03
−2.73***
-
-
3
0
0.48
1.38
0
0.44
22
Sierra Leone
SLE
1960−2018
59
2
0
−1.94
−2.05**
-
-
1
0
−4.32
4.03
0
−1.27
23
Seychelles
SYC
1960−2018
59
3
0
0.70
−2.14**
-
-
3
0
−1.83
−1.79*
-
-
24
Chad
TCD
1960−2018
59
1
0
−0.85
−2.28**
-
-
1
0
−0.77
−1.00
0
−1.48
25
Togo
TGO
1960−2018
59
1
1
−3.88
2.28
1
−2.94**
1
1
−5.62
4.17
1
−2.93
26
South Africa
ZAF
1960−2018
59
2
1
−2.34
−1.54*
1
−1.75
2
1
−2.82
−1.88**
-
-
27
Zambia
ZMB
1960−2018
59
5
2
−3.11
2.10
2
−3.06**
5
2
−1.77
1.96
2
−1.63
28
Zimbabwe
ZWE
1960−2018
59
5
2
−3.86***
−2.63**
-
-
5
2
−3.46**
−2.48**
-
-
29
Gambia, The
GMB
1966−2018
53
5
0
−2.42
1.62
0
−2.38
5
0
−2.19
1.61
3
−4.30***
30
Mali
MLI
1967−2018
52
4
0
0.12
−2.46**
-
-
4
0
−1.81
−1.90**
-
-
31
Guinea−Bissau
GNB
1970−2018
49
5
0
−2.56
−0.90
0
−2.65*
1
0
−2.47
1.05
0
−2.56
32
Eswatini
SWZ
1970−2018
49
3
0
1.19
2.42
0
0.58
3
0
−1.68
2.06
0
−2.15
33
Mauritius
MUS
1976−2018
43
4
2
3.79
4.09
3
4.09
4
2
2.10
4.37
3
0.97
34
Angola
AGO
1980−2018
39
4
1
−2.36
−2.70**
-
-
1
2
1.33
−2.74***
-
-
35
Comoros
COM
1980−2018
39
4
1
−3.02
2.01
1
−2.08
4
1
−1.87
1.87
2
−1.64
36
Cabo Verde
CPV
1980−2018
39
3
1
−0.27
2.18
1
−0.47
3
1
−2.19
1.75
1
−2.65
37
Equatorial Guinea
GNQ
1980−2018
39
1
2
−6.33***
−6.77***
-
-
4
2
−3.38**
−1.72*
-
-
38
Mozambique
MOZ
1980−2018
39
1
1
1.28
−3.43***
-
-
1
2
−0.48
−1.41*
-
-
39
Namibia
NAM
1980−2018
39
1
1
0.19
−3.42***
-
-
4
0
−2.46
−1.80**
-
-
40
Ethiopia
ETH
1981−2018
38
1
0
3.36
−3.42***
-
-
1
0
2.30
−1.77**
-
-
41
Uganda
UGA
1982−2018
37
1
1
−1.24
−2.45**
-
-
4
1
−2.57
−0.86
1
−2.58
42
Guinea
GIN
1986−2018
33
1
2
8.15
6.22
2
1.58
2
0
2.78
−3.79***
-
-
43
Tanzania
TZA
1988−2018
31
5
2
0.62
−1.29*
2
0.86
2
1
−5.04
5.04
1
−5.48***
Note: k = frequency; p = optimal lag length chosen via SIC; ***, ** and * denote significance at 1%, 5% and 10% respectively. WADF is used for the cases where t-statistic is found to be insignificant.
Region
Total countries
Selected countries
Stationary under only "Constant" specification(A)
Stationary under only "Constant & trend" specification (B)
Stationary under both specifications(A∩B)
Stationary either specification(AUB)
Total stationary
Total Non-Stationary
East Asia and Pacific
38
26
3
4
1
6
6
20
Europe and Central Asia
58
46
7
21
3
25
25
21
Latin America and the Caribbean
42
36
7
12
5
14
14
22
Middle East and North Africa
21
17
7
9
4
12
12
5
North America
3
3
0
1
0
1
1
2
South Asia
8
6
0
1
0
1
1
5
Sub-Saharan Africa
48
43
6
7
2
11
11
32
Total
218
177
30
55
15
70
70
107
Figure 1. Flowchart showing the systems biology approach adopted in this study for identifying the potential candidate genes shared between diabetes and obesity
Figure 2. Differential expression analysis of protein coding genes. A) Volcano plot of the genes that differ in expression between obesity and lean samples. B) Volcano plot of genes that differ in expression between T2D and lean samples. C) Number of DEGs and their up and down regulated status in obese and T2D samples. D) Shared DEGs between Obesity and T2D samples
Figure 3. PPI Network parameters and their correlation. A) Distribution of association score among protein-protein interactions. B) Protein Interaction Network with 1203 nodes and 3211 edges. C) Correlation between centrality parameters betweenness and closeness. D) Correlation between centrality parameters Eigen vector and closeness
Figure 4. The bar and dot graph of up- and down- DEGs A) The relationship between obesity and T2D-related genes that are significantly differentially expressed. B) The number of influential genes and their shared association with obesity and T2D as measured by topological parameters such as degree, betweenness, closeness, and eigen vector
Figure 5. Heatmap of differentially expressed genes shared between T2D and Obesity representing similar expression pattern among obesity and T2D when compared to lean samples
Figure 6. Overview of functional enrichment. A) Topological parameters were used to identify total ontologies among the filtered genes. B) Total ontologies identified among obesity and T2D shared genes. C) A summary of gene ontologies enriched in biological processes, molecular functions, and cellular components
Figure 7. The number of modules generated based on the enrichment analysis in each tissue, where A) adipose, and B) pancreatic tissue
Figure 8. The GWAS trait network of Q1, and Q2, based on the topological parameters. Nodes from the Q1 and Q2 categoriesare highlighted in red color