Original Articles
Recent effective population size estimated from segments of identity by descent in the Lithuanian population
2017 Volume 125 Issue 2 Pages 53-58
Details
2017 Volume 125 Issue 2 Pages 53-58
The effective population size (Ne) quantifies the rate at which genetic diversity is eroded by genetic drift, a fundamental process of evolutionary change, and also provides an insight into the demographic history and dynamics of modern human populations. The main interest of this study was to reconstruct the recent effective population size, by using inferred long segments of identity by descent (IBD), and to estimate the effective/census size ratio in the Lithuanian population. We used Illumina 770K HumanOmniExpress-12v1.0 array data of 295 unrelated individuals of the Lithuanian population. IBDseq v. r1206 and IBDNe v. 04Sep15.e78 software packages were used to detect IBD segments and to estimate Ne, respectively. We estimated the effective population size in Lithuania 50 generations (g) ago to be 11900, whereas for g = 0 (1991) the effective population was 417000 (95% confidence interval, CI [218000; 1150000], and the census size was 3701968. We evaluated the ratio of effective size to census (N) size. The estimates of Ne were approximately one-tenth of the census size. We conclude that natural levels of fluctuations in the Lithuanian population size probably caused the small values of Ne/N. Because of extrapolation of slowing growth rates and migration of the Lithuanian population, this estimate might be correct, as the census size is expected to be several times larger than the Ne.
The effective population size is an important population parameter that helps to explain the evolution and expansion of human populations. By definition Ne is a measure of the number of independent breeding individuals in an ideal population and is much lower than the actual census size (N) (Wright, 1931; Charlesworth, 2009). According to Felsenstein (1971), the effective population size in modern human populations is supposed to be around one-third of the census size, though other estimates are much lower. Both variables, Ne and N, play key roles in determining the extent to which populations can avoid extinction due to demographic, environmental, or genetic stochastic events, such as reduction in genetic diversity and genetic drift in small populations or temporary recruitment failures and environmental catastrophes (Soule, 1987; Boyce, 1992). Furthermore, simple conversions between Ne and N would save much time and money as the estimation of one variable could be used to infer the other (Lukart et al., 2010).
Nowadays we have numerous measures to estimate Ne, however, few existing methods are capable of detecting recent demographic events as most are focused on ancient population sizes, i.e. hundreds to thousands of generations before the present (Palamara et al., 2012). In this study, we aimed to reconstruct the recent effective population size using the latest non-parametric approach based on long segments of identity by descent (IBD) (Browning and Browning, 2015) and to calculate the effective/census size ratio values in the Lithuanian population.
The data set consisted of 295 samples from unrelated Lithuanian individuals. The average age of participants was 53 years. The samples were collected randomly from six ethnolinguistic groups of Lithuania: three groups of Aukštaičiai (western (n = 52), southern (n = 51), and eastern (n = 48)) and three groups of Žemaičiai (northern (n = 61), western (n = 24), and southern (n = 59)) (Figure 1). We excluded full siblings and closer relatives from the study based on information from the questionnaires about every individual’s relationships.
Map of Lithuanian ethnolinguistic groups.
Genomic DNA was extracted from whole venous blood using either the phenol–chloroform extraction method or automated DNA extraction (TECAN Freedom EVO, TECAN Group Ltd, Männedorf, Switzerland) based on the paramagnetic particle method. DNA concentration and quality were measured by a NanoDropR ND-1000 spectrophotometer (NanoDrop Technologies Inc., USA).
Single-nucleotide polymorphism (SNP) genotyping of 295 samples was performed with an Illumina 770K HumanOmniExpress-12v1.1 array (Illumina, San Diego, CA, USA) at the Department of Human and Medical Genetics, Faculty of Medicine, Vilnius University, Lithuania using the standard Illumina Infinium® HD Assay Ultra protocol recommended by the manufacturer (catalog no. WG-901- 4005). Genotyping data quality control was performed according to the manufacturer’s standard recommendations.
Individuals with call rates <98% and standard deviations (SD) of log R ratio >0.3 were excluded from further analysis. Individuals or SNPs with >10% missing data, minor allele frequency (MAF) <0.01 and Hardy–Weinberg equilibrium (HWE) test P-values of <10−4 were excluded. After quality control we were left with 1 of 295 individuals removed for low genotyping (MIND > 0.1) and after frequency and genotyping pruning, 568040 SNPs remained.
This study is part of the LITGEN project, which was approved by the Vilnius Regional Research Ethics Committee 235 No. 158200-05-329-79, date: 2011-05-03. Written informed consent was received from all of the study participants.
Ne estimationTo infer the history of the recent effective population size we used a non-parametric method, based on the Wright–Fisher discrete-generation model, implemented in the open-source IBDNe v. 04Sep15.e78 software package, published by Browning and Browning (2015). This method is based on segments of IBD that provide information about Ne around 50 generations from the present using SNP array data. The distribution of detected IBD lengths is presented in Additional File 2. The length filter used to detect IBD segments with the IBDseq v. r1206 software package was 7 cM.
We estimated the recent effective population size for 50 generations (g), or 1250 years, from the present, from IBD segments by using the newest non-parametric approach (Browning and Browning, 2015) on the Lithuanian population (Figure 2) (Appendix Table 1). One generation is considered to be 25 years. Fifty generations ago the effective population size in Lithuania was 11900, whereas from 2015 to 1991, corresponding to generation 0, it was 417000 (95% confidence interval, CI [218000; 1150000]), and the mean census size was 3373154. The average estimated Ne for generations 30–50 was 16228. Increased exponential growth is observed from generation 25.
Recent effective population size estimated in the Lithuanian population for 50 generations with 95% confidence intervals.
We evaluated the ratio of effective population size to census size, Ne/N, in the Lithuanian population at selected time points. We obtained reliable census size data for only three generations of the Lithuanian population from Eurostat (Appendix Table 2), the statistical yearbook of Lithuania (Official Statistics Portal: http://osp.stat.gov.lt/en/statistikos-leidiniu-katalogas?%20publication=1673) and from the ebook of the first Population Census of pre-war Lithuania (Stulginskis and Galvanauskas, 1923). For generation 1 (1990–1966), the Ne was 383000 (95% confidence interval, CI [217000; 938000]), and the mean census size was 3351015. Furthermore, for generation 2 (1965–1941), the Ne was 352000 (95% confidence interval, CI [217000; 763000]), and the mean census size was 2816800.
The estimated ratio was 0.125 (95% CI [0.077; 0.271]) for g = 2 (corresponding to 1941), 0.114 (95% CI [0.065; 0.280]) for g = 1 (corresponding to 1966), and 0.124 (95% CI [0.065; 0.341]) for g = 0 (corresponding to 1991) (Figure 3). The estimates of Ne were approximately one-tenth of the Lithuanian population size based on the census.
The ratio between Ne/N in the Lithuanian population for three generations with 95% confidence intervals.
There are two important parameters to be estimated in natural populations: the effective population size, Ne, and the census population, N. Knowing the ratio of Ne/N is useful for estimating the effective size of a population from census data and for examining how different ecological factors influence effective size (Kalinowski and Waples, 2002). The main interest of this study was to estimate the recent effective population size by using IBD segments and to relate this to census size in the Lithuanian population. The obtained results for 50 generations, assuming a 25 year generation time, show exponential growth of the population, with increased growth rates from generation 25.
Estimated effective/census size ratio values, for the first three generations in the Lithuanian population, were very similar, at around 0.1. The obtained values of Ne/N ratios are small (0.1) compared with other genetics-based estimates of between 0.21 and 0.65 (Frankham, 1995). According to Nunney and Campbell (1993) and Nunney (1996), the Ne/N ratio is usually close to 0.5 and only rarely outside the range 0.25–0.75. However, very low estimates of Ne/N (<0.1) raise the possibility that other factors acting to reduce Ne have been underestimated, e.g. variation in female fecundity (Nunney, 1996). Furthermore, natural levels of fluctuations in population size, in the hypothetical absence of any other influence, are often sufficient to depress Ne/N to small values (Vucetich et al., 1997). A single factor is sufficient to produce very small Ne/N values and additional factors tend to depress these values even further (Vucetich et al., 1997).
We conclude that natural levels of fluctuations such as variance in size, reproduction, sex ratio, and the degree to which generations overlap probably caused the small values of Ne/N in the Lithuanian population. The population of Lithuania is small and historically suffered effects of population bottlenecks (a rapid decrease in population size in generation 2) and expansions (a rapid increase in population size in generation 0) which might produce very small Ne/N values. Furthermore, the small sample size of this study and Lithuanian population structure could introduce oscillations for the most recent generations. However, considering our results we think that the true effective size is contained within the bootstrap confidence interval. Browning and Browning (2015) state that the results for 200 individuals are quite good, although with 100 or 200 individuals there is less precision and more oscillation for the most recent generations when analysing 1000 individuals.
This study was approved by the Vilnius Regional Research Ethics Committee No. 158200-05-329-79, date: 2011-05-03. Written informed consent was received from all of the study participants.
Competing InterestsThe authors declare no competing interests.
This study is part of the LITGEN project, which was approved by the Vilnius Regional Research Ethics Committee No. 158200-05-329-79, date: 2011-05-03. We also thank all investigators (L. Ambroazitytė, I. Uktverytė, I. Domarkienė, R. Meškienė) who contributed to the sample genotyping.
Histogram of detected IBD lengths shared by pairs of individuals.
| Generation | Lower 95%CI | Ne | Upper 95%CI |
|---|---|---|---|
| 0 | 218000.00 | 417000.00 | 1150000.00 |
| 1 | 217000.00 | 383000.00 | 938000.00 |
| 2 | 217000.00 | 352000.00 | 763000.00 |
| 3 | 214000.00 | 323000.00 | 614000.00 |
| 4 | 209000.00 | 296000.00 | 488000.00 |
| 5 | 200000.00 | 271000.00 | 382000.00 |
| 6 | 192000.00 | 248000.00 | 346000.00 |
| 7 | 181000.00 | 227000.00 | 294000.00 |
| 8 | 171000.00 | 208000.00 | 273000.00 |
| 9 | 153000.00 | 190000.00 | 253000.00 |
| 10 | 144000.00 | 172000.00 | 230000.00 |
| 11 | 124000.00 | 155000.00 | 202000.00 |
| 12 | 106000.00 | 140000.00 | 173000.00 |
| 13 | 93400.00 | 125000.00 | 154000.00 |
| 14 | 84800.00 | 111000.00 | 137000.00 |
| 15 | 77700.00 | 98400.00 | 122000.00 |
| 16 | 71500.00 | 87000.00 | 107000.00 |
| 17 | 65700.00 | 77100.00 | 94600.00 |
| 18 | 58100.00 | 68800.00 | 84500.00 |
| 19 | 50300.00 | 61500.00 | 74900.00 |
| 20 | 44100.00 | 55000.00 | 65900.00 |
| 21 | 38800.00 | 49300.00 | 58600.00 |
| 22 | 35300.00 | 44900.00 | 52200.00 |
| 23 | 32800.00 | 40900.00 | 47800.00 |
| 24 | 30200.00 | 37500.00 | 43600.00 |
| 25 | 28000.00 | 34300.00 | 40300.00 |
| 26 | 26000.00 | 31300.00 | 37500.00 |
| 27 | 24200.00 | 28700.00 | 34900.00 |
| 28 | 23000.00 | 26900.00 | 32600.00 |
| 29 | 21800.00 | 25300.00 | 30500.00 |
| 30 | 20600.00 | 23700.00 | 29100.00 |
| 31 | 19600.00 | 22300.00 | 27500.00 |
| 32 | 18800.00 | 21000.00 | 25800.00 |
| 33 | 17900.00 | 20000.00 | 24400.00 |
| 34 | 17200.00 | 19100.00 | 23900.00 |
| 35 | 16400.00 | 18500.00 | 23000.00 |
| 36 | 15700.00 | 17800.00 | 22000.00 |
| 37 | 15100.00 | 17100.00 | 21200.00 |
| 38 | 14600.00 | 16700.00 | 20400.00 |
| 39 | 14100.00 | 16100.00 | 19400.00 |
| 40 | 13500.00 | 15700.00 | 18700.00 |
| 41 | 12900.00 | 15400.00 | 18000.00 |
| 42 | 12300.00 | 14900.00 | 17900.00 |
| 43 | 11800.00 | 14300.00 | 18000.00 |
| 44 | 11400.00 | 13600.00 | 17400.00 |
| 45 | 11000.00 | 13200.00 | 17100.00 |
| 46 | 10300.00 | 12700.00 | 16600.00 |
| 47 | 9960.00 | 12600.00 | 16300.00 |
| 48 | 9690.00 | 12200.00 | 16000.00 |
| 49 | 9430.00 | 12000.00 | 15700.00 |
| 50 | 9280.00 | 11900.00 | 15300.00 |
| year | population |
|---|---|
| 1960 | 2755600 |
| 1961 | 2801500 |
| 1962 | 2845600 |
| 1963 | 2881100 |
| 1964 | 2916800 |
| 1965 | 2953600 |
| 1966 | 2989300 |
| 1967 | 3026800 |
| 1968 | 3062000 |
| 1969 | 3095700 |
| 1970 | 3118941 |
| 1971 | 3160437 |
| 1972 | 3197645 |
| 1973 | 3229598 |
| 1974 | 3259277 |
| 1975 | 3288510 |
| 1976 | 3314794 |
| 1977 | 3342533 |
| 1978 | 3367538 |
| 1979 | 3391490 |
| 1980 | 3404194 |
| 1981 | 3422210 |
| 1982 | 3443684 |
| 1983 | 3470673 |
| 1984 | 3499711 |
| 1985 | 3528698 |
| 1986 | 3560388 |
| 1987 | 3597439 |
| 1988 | 3635295 |
| 1989 | 3674802 |
| 1990 | 3693708 |
| 1991 | 3701968 |
| 1992 | 3706299 |
| 1993 | 3693929 |
| 1994 | 3671296 |
| 1995 | 3642991 |
| 1996 | 3615212 |
| 1997 | 3588013 |
| 1998 | 3562261 |
| 1999 | 3536401 |
| 2000 | 3512074 |
| 2001 | 3486998 |
| 2002 | 3454637 |
| 2003 | 3431497 |
| 2004 | 3398929 |
| 2005 | 3355220 |
| 2006 | 3289835 |
| 2007 | 3249983 |
| 2008 | 3212605 |
| 2009 | 3183856 |
| 2010 | 3141976 |
| 2011 | 3052588 |
| 2012 | 3003641 |
| 2013 | 2971905 |
| 2014 | 2943472 |
| 2015 | 2921262 |
Source of data Eurostat
Last update 03.03.16
Extracted on 08.04.16
Sex-pooled
