MHC genotyping (high-throughput sequencing)

To get a better resolution of MHC polymorphism and the mechanisms that may contribute to its maintenance (recombination and selection) in the Eurasian coot, we genotyped key MHC regions (selected PBR-coding exons) in all captured individuals (283 individuals genotyped at MHC-I, 230 genotyped at MHC-II). Population genotyping focused on a single exon per MHC class (exon 3 at MHC-I and exon 2β at MHC-II), as these exons are traditionally targeted in avian MHC research (allowing direct comparisons across species) and their polymorphism is expected to be well representative for the entire PBR region³⁴. To genotype MHC-I exon 3, we used primers MHCI-int2F (5’-CATTTCCCTYGTGTTTCAGG-3’) and MHCI-ex4R (3’-GGGTAGAAGCCGTGAGCRC-5’), which were originally designed for accipitrid birds³⁵. Primer MHCI-int2F binds to the conserved flanking region of intron 2 and primer MHCI-ex4R binds to the conserved region of exon 4. Specificity of these primers towards coot MHC-I genes was verified using our genome assembly, showing no mismatches within the 3-terminus region, which is crucial for effective PCR amplifications³⁶. Consequently, non-specific MHC-I amplifications (allele drop out) were unlikely. The length of the entire amplicon was 411 bp, including almost entire exon 3 (273 bp out of 276 bp). Species-specific primers Fuat-Ex2Fw (5′-CTGACCRGCCTCCCTGCA-3′) and Fuat-Ex2Rv (5′-TTGTGCCAYACACCCACC-3′) were used to amplify MHC-II. These two primers were originally designed for the Eurasian coot²⁶ and they successfully amplify the entire MHC-II exon 2 (270 bp), binding to the flanking regions of intron 1 and 2. In each PCR reaction we used fusion primers with Illumina Nextera Transposase adapter sequences (Illumina Corp., San Diego, CA, USA) and 7-bp barcodes to identify the samples. PCR amplifications were carried out in a final volume of 20 μl containing 20–80 ng genomic DNA (1 µl of DNA isolate), 10 µl of 2X HotStarTaq Plus MasterMix Kit (Qiagen, Venlo, The Netherlands), 8 µl of deionized water and 0.5 µl of each primer. PCR protocols followed Alcaide et al.³⁵ for MHC-I and Alcaide et al.²⁶ for MHC-II, although in both cases the number of PCR cycles was reduced to 25 to suppress the formation of artificial chimeras, which could confound the correct interpretation of Illumina sequencing results. The effects of PCR reactions were confirmed for each sample by visual examination of band intensities on 2% agarose gel electrophoresis. To purify PCR products we used AMPure XP magnetic beads (Beckman Coulter, Brea, CA, USA) and concentration estimates were quantified using Quant-iT PicoGreen dsDNA marking kit (Thermo FisherScientific, Waltham, MA, USA). Separate libraries for MHC class I and II were prepared using equimolar concentrations of purified PCR products and NEB-Next DNA Library Prep Master Mix Set for Illumina (New England Biolabs, Ipswich, MA, USA). Both libraries were sequenced on the 2 × 250 bp Illumina MiSeq platform.

In the processing of raw Illumina data we used an online webserver, the Amplicon Sequencing Analysis Tools (AmlpiSAT)³⁷, and followed recommendations by Biedrzycka et al.³⁸. In the first step we used the Amplicon Sequencing MERGing (AmpliMERGE) tool, which merges paired-end reads, optimizing their overlapping lengths according to amplicon data³⁹. Next, we used the Amplicon Sequencing Assignment (AmpliSAS) tool, which performs read demultplexing, variant clustering and putative allele filtering based on user-specified criteria. For the clustering step (identification of reads resulting from genotyping errors and clustering them with reads identified as true alleles) we used default AmpliSAS settings for Illumina data, including a substitution error rate of 1%, an indel error rate of 0.001% and the minimum dominant frequency of 25%. Finally, we used AmpliSAS to filter for clusters that are likely to be artefacts, including chimeras and other low-frequency artefacts (>3%) that were retained through the clustering step. Samples with amplicon depth of less than 300 reads were excluded from the analyses and the maximum amplicon depth was, by default, set to 5000 reads because of AmpliSAS performance reasons. The average amplicon depth prior to the processing was 4453 ± 66 [SE] reads for MHC-I and 2616 ± 97 [SE] reads for MHC-II. We obtained validated MHC-I and MHC-II genotypes for 270 and 220 individuals, respectively. Technical reproducibility of validated sequences was 93.7%, as estimated using 36 technical replicates (i.e. samples for which two amplicons were obtained in independent PCR reactions and sequenced). To align all unique MHC class I and II sequences we used Geneious v10.0.5 (Biomatters Ltd., Auckland, New Zealand). We removed intron regions from the alignments and we inferred alleles based on the exon fragments only.