Parasite-mediated selection is widespread at loci involved in immune defence, but different defences may experience different selective regimes. For defences involved in clearing infections, purifying selection favouring a single most efficacious allele likely predominates. However, for defences involved in sensing and recognizing infections, evolutionary arms races may make positive selection particularly important. This could manifest primarily within populations (e.g., balancing selection maintaining variation) or among them (e.g., spatially varying selection enhancing population differences in allele frequencies). We genotyped three toll-like receptors (TLR; involved in sensing infections) and three avian beta-defensins (involved in clearing infections) in 96 song sparrows (Melospiza melodia) from three breeding populations that differ in disease resistance. Variation-based indicators of selection (proportion of variable sites, proportion of nonsynonymous SNPs, proportion of sites bearing signatures of positive or purifying selection, rare allele frequencies) did not differ appreciably between the two locus types. However, differentiation was generally higher at infection-sensing than infection-clearing loci. Allele frequencies differed markedly at TLR3, driven by a variant predicted to alter protein function. Geographically structured variants at infection-sensing loci may reflect local adaptation to spatially heterogeneous parasite communities. Selective regimes experienced by infection-sensing versus infection-clearing loci may differ primarily due to parasite-mediated population differentiation.

Methods are described in Boyd RJ, MR Denommé, LA Grieves, and EA MacDougall-Shackleton. Stronger population differentiation at infection-sensing than infection-clearing innate immune loci in songbirds: different selective regimes for different defences. Evolution (in press).

References in datafiles to populations QUBS, RARE and AFAR correspond respectively to eastern, central and western locations described in the paper.

Dataset includes:

Summary file with sample ID, population, and whether or not sequence was obtained from each of six loci. Each row represents an individual song sparrow (n=96) and columns correspond to the six loci sequenced. This file is named Boyd2Post_PopulationsAndLocusCoverage.

5 DNA sequence files (one for each variable locus) presented in FASTA format. For each file, the first sequence is a reference sequence from GenBank, and the rest are song sparrow sequences used in this study. File names are as follows,

Boyd2Post_AvBD2sequences.txt (61 sequences), 

Boyd2Post_AvBD7sequences.txt (75 sequences), 

Boyd2Post_TLR1LBsequences.txt (73 sequences), 

Boyd2Post_TLR3sequences.txt (72 sequences), 

Boyd2Post_TLR4sequences.txt (71 sequences).

Datafile of single-nucleotide polymorphisms (SNPs) detected. The file is named Boyd2Post_SNPs_Fst. Each row represents a SNP. Columns are as follows,

Locus: the locus where the SNP occurs

AA site: amino acid position, relative to the start of the sequence

Mutation type: synonymous or nonsynonymous

Fst: observed population structuring across the 3 populations, as calculated in Genepop

Phased haplotype datafile. The file is named Boyd2Post_PhasedHaplotypes. Each row represents an individual song sparrow (n=85) and columns correspond to the five variable loci sequenced.