The effects of artificial selection on genetic variation of some immune genes in Gallus gallus

To research effects of the artificial selection of Gallus gallus on G. domesticus' nucleotide diversity of immune genes, sequence polymorphisms of G. domesticus (23 genes), G. gallus (23 genes), G. lafayetti (17 genes), and G. sonneratii (17 genes) were obtained from GenBank. The data set included 819 polymorphisms. Immune gene polymorphism and selection efficiency in the data from those four species of Gallus were calculated. By calculating the qw (Watterson's estimator) of each site, an average qw for each species and the minimum number of re-combinations in each species and by estimating the selection efficiency for G. domesticus and G. gallus, neither significant nucleotide diversity nor genetic-diversity-qwdifference was found between G. domesticus and G. gallus. The results indicated that the patterns of genetic diversity in G. domesticus were strongly influenced by recombination and, because Tajima's D has a negative value, recombination was the main mechanism responsible for the immune gene evolution of G. gallus.


Introduction
Domestic animals have often been artificially selected for certain traits over several thousand years.Poultry domestication is the genetic modification of a wild species to create a new form of a bird to meet human needs.Improvement after domestication has also resulted in striking changes in yield, immune system, biochemical composition and other traits.
The domestic chicken is descended primarily from the Red Junglefowl (G.gallus) in Southeast Asia nearly 10 000 years ago (Crawford 1990).But at least one other species must have contributed, specifically the grey jungle fowl (Wong et al. 2004), to the domestic chicken.The size, shape and production of the modern domesticated chicken have been sculpted by artificial selection for at least 2 000 years, likely contain many important lessons about the genetic architecture of phenotypic variation and the mechanistic basis of selection.Indeed, chicken and other domesticated species played an important role in Darwin's »On the Origin of the Species«, as they provided vivid examples of descent with modification.
Most domesticated animals have experienced »a domestication bottlenec« that reduced genetic diversity relative to their wild ancestor (Buckler 2001).This bottleneck affects all genes in the genome and modifies the distribution of genetic variation among loci.Selection is similar to a more severe bottleneck (Galtier 2000) that removes most of the genetic variation from a target locus.Chicken (G.domesticus) showed a high density of SNP and a high recombination rate, which made it possible to perform high-throughput genotyping to evaluate the existing genetic diversity in chicken at the genome level compared to other species.However, relatively little progress has been made on systematically identifying which immune gene sites of G. domesticus genome were influenced by selective breeding during the natural history of chicken.
Here, genetic variations of 23 gene fragments in a sample of Gallus Genus 4 species, G. domesticus, G. gallus, G. lafayetii, G. sonneratii on the basis of gene sequence polymorphism were reported and the effects of artificial selection on some immune genes were analysed in G. gallus.The multi-locus analysis is a powerful way to detect adaptation at the population level, so that comparative researches of diversity and recombination in the Gallus genus would help us to comprehensively understand the immune genetic structure and selection in domestication.

Gallus families sequence polymorphism data set
A total of 819 data sets of Gallus genus gene polymorphisms (1 159 to 9 398 base pairs) was obtained from Popset of GenBank and Daniel G. Bradley (2010) including sequence poly morphisms of G. domesticus 23 genes, G. gallus 23 genes, G. lafayetii 17 genes and G. sonneratii 17 genes.Each group was aligned by eye using CLUSTALW (Thompson et al. 1994).Alignments of all groups are available on request.

Polymorphism sequence data analyses
The average variability Pi and minimum number of the recombination parameter were cal culat ed by using DnaSP v. 5 (Librado & Rozas 2009).Insertions/deletions (indels) were ex clud ed from all estimates.To investigate the evidence of the non-neutral evolution, the D test of Tajima was applied (Tajima 1989).
The diversity was measured by Watterson's estimator of the population mutation pa rameters (q w ), which was calculated separately from the non-synonymous and silent sites for the con-specific gene fragments.The parameter represents the per-site diversity.
P is the number of synonymous polymorphisms, L is the number of synonymous sites and n is the number of the sequence sampled.
We calculated the efficiency of selection for four species.Measurement of selection efficiency is as following: -----------------(2) where P n , P s and P i are the numbers of non-synonymous, synonymous and intron poly morphisms.L n , L s and L i are the numbers of non-synonymous, synonymous and intron sites for each gene in each species.q n is for non-synonymous sites.q s+i is for synonymous and intron correspondingly.q were arc-sine transformed.Recombination parameter was log x+1 transformed.After the calculation of q s for synonymous sites and q i for intron sites, the weighted average of q s or/and q i from different genes for same species was made.

Whole sequence segment variations in G. domesticus and G. gallus
More than 84 kb of the DNA sequence was obtained across 23 immune genes.Sequence data for the CR1 and OTC gene were obtained from only G. domesticus and G. gallus.Summary statistics of the number of segregating sites, nucleotide diversity Pi, q w , Tajima's D test and the minimum number of recombination events were shown in Table 1.Single variable regressions of immune gene nucleotide diversity showed that no significant difference was found between G. domesticus and G. gallus (F 2,23 =1.39,P=0.5845>0.05).Neither did the diversity-q w of the fragments (F 2,22 =0.074, P=0.7866>0.05).
The population recombination parameter, ρ, is the other key parameter in simple population genetic models.However, the estimation of ρ requires considerably larger segments of contiguous DNA to be sequenced (Hudson 2001).The relatively short sequences obtained in this study are not sufficient to provide reliable locus-specific estimates of ρ.Instead, the estimation of the minimum number of the recombination parameter was obtained from the two sample species.Single variable regressions revealed that the recombination parameter of G. domesticus was significantly higher than that of G. gallus (F 2,23 =6.160, P=0.0169<0.05) after recombination parameter was log x+1 transformed.By using DnaSP, recombination parameter was estimated, which is inversely proportional to LD (linkage disequilibrium).The average of estimates of recombination parameter in G. domesticus is 164 % of that in G. gallus, while the average of estimates of q w in G. domesticus is 95 % of that in G. gallus (Figure 1).Thus, the recombination parameter in G. domesticus has been reduced more drastically than the population mutation parameter q w , contrary to what has been expected under a population Different sites' q w , an average q w and efficiency of selection for 4 species G. domesticus G. gallus G. lafayetii G. sonneratii Gene q s q n q s+i q s q n q s+i q s q n q s+i q s q n q s+i KK34 0.

Rm(d)
The first row illustrates the relationship between mean values of q w in G. domesticus (y-axis) versus G. gallus (x-axis).Dashed diagonal lines have a slope of 1.0, representing equal diversity between taxa; solid lines lines are regression lines.Each square represents a single gene.The second row plots the relationship between estimates of minimum number of recombination in G. domesticus (y-axis) versus G. gallus (x-axis).
Figure 1 Patterns of diversity in G. domesticus and G. gallus at 23 gene fragments bottleneck (Wall et al. 2002).The results suggest that patterns of immune gene diversity in G. domesticus are strongly influenced by recombination.
Tajima's D test is to distinguish between a DNA sequence evolving neutrally and DNA evolv ing under a non-random process, including selection, demographic expansion or contrac tion.In order to perform the test, homologous DNA for at least three individuals was required, so that five sequence polymorphism groups were not available (Table 1).In terms of 23 sampled genes, TLR5 in G. domesticus was found to be statistically significant for Tajima's D (P<0.05).In principle, this could potentially indicate a deviation from neutrality, possibly due to strong selection.An average negative Tajima's D in G. domesticus (−0.023), in contrast to G. gallus (0.036), signifies slightly more low frequency polymorphisms, indicating a population size expansion and/or selection in G. domesticus.

Variation of different sites of gene sequence and efficiency of selection in Gallus genus
Across the 23 genes, 84 013 base pairs were aligned, including 831 mutation sites.The com par i son of different sites' q w , a weighted average of q w and selection efficiency for four species was shown in Table 2.The weighted average of q n for non-synonymous sites was arranged as follows: G. domesticus (0.00173), G. gallus (0.00171), G. lafayetii (0.00117) and G. sonneratii (0.00094).In the top-to-bottom order, while the difference of the weighted average q s+i for four species was not found, the arrangement was as follows: G. domesticus G. gallus (0.00293), G. lafayetii (0.00216) and G. sonneratii (0.00290), respectively.q n /q w represented the percent of mutations of the non-synonymous in total gene segment mutations.G. domesticus showed the highest value of q n /q w (0.60192) for four species an alysed and suggested that G. domesticus' immune genes may undergo more strong pres sure of selection.

Discussion
In this paper, sequence segment variations, diverse sites weighted mutations (q w ), re combi na tion parameter and efficiency of selection within Gallus genus were determined.The results demonstrated the artificial selection of G. domesticus and indistinguishable immune genetic diversity with three other species.
A diversity index (Pi) analysis indicated that the overall nucleotide variability of all 23 im mune genes for G. domesticus and G. gallus were approximately 0.003 and showed no dif fer ence, as mentioned above.The population recombination rate ρ is a fundamental pa ram eter for evolutionary biology.Not only recombination is a key force shaping the archi tec ture of genomes, but also distribution across genomic regions is essential for as soci a tion studies of traits.However, the estimation of the population recombination rate is not an easy task.Adequate and reliable locus-specific estimates of ρ could not be provided with relatively short sequences, as we failed in estimation by LDhat v2.0 (McVean 2004), a package for the analysis of recombination rates from population genetic data.We turned to calculate the minimum number of the recombination parameter.The result revealed that the recombination parameter in G. domesticus was higher than that in G. gallus in most of the sampled loci (21/23).
Of 84 013 base pairs, 831 mutation sites were found.This is somewhat more than the ex tensive sequence diversity present in domestic chicken (~5 single nucleotide polymorphisms per kilobase in pairwise comparisons) (Wong et al. 2004), mainly as a result of the sample size.
Regions of intergenic, noncoding DNA where levels of variation are expected to be higher (Zwick et al. 2000) may provide a different picture of diversity.Our estimates of the weighted average of q w for both silent mutation sites and non-synonymous sites indicated that q w for silent mutation sites in total species were higher than that for non-synonymous sites, as we expected.G. lafayetii has the lowest synonymous mutation q w =0.00216 among four species, which is obviously correlated to its effective population size.
The mutation parameter of non-synonymous sites-q n -and percent of non-synonymous mutations in total segment mutations-q n /q w -for G. domesticus and G. gallus were higher than that for G. lafayetii and G. sonneratii, which indicated that the immune genes of G. domesticus and G. gallus could undergo a stronger directional selection pressure.This was inconsistent with previous researches that selection for body weight in chicken has depressed immune performance (Miller et al. 1992) and antibody production (Cheema et al. 2003).Now that almost equivalent q w for G. domesticus and G. gallus, there should be some factors which could decrease nuclear diversity of G. domesticus since this species obviously have high effective population size, Ne.Of these factors, high recombination event in G. domesticus was inferred to be essential.
The highest efficiency of selection (0.60192) was found in G. domesticus (Table 2).This selection was mainly described as artificial selection for the needs of human being here, which could be confirmed by negative noticeable Tajima's D.
A population bottleneck was not found by analysis of mutation and recombination pa rameter.Nevertheless, G. domesticus did experience severe population bottleneck (Mason 1984) although this bottleneck effect did not result in a substantial loss of genetic diversity.Abroad crossing between breeds and higher recombination could be fundamental explanations for undifferentiated diversity between G. domesticus and G. gallus.
Of particular interest will be to define the number of loci responsible for shaping the diversity of form and function, the types of genes and genetic variation therein that have responded to artificial selection.Although our results did not provide definitive answers to these issues, they did afford some insight into the mechanistic basis of artificial selection.Despite the insights gleaned from our data, one limitation of this study was that it did not provide information about more gene sequence polymorphisms.The difference in part reflected differences in sampling.

Table 1
Pi, Theta, TajimaD and recombination parameter of 23 gene sequence for 4 species The effects of artificial selection on genetic variation of some immune genes in Gallus gallus G. d.: Gallus domestiucs, G. g.: Gallus gallus, L: length of gene, Ns: number of segregating sites, D: Tajima's D, Rm: recombination parameter, na: not available Zhang et al.: