Genome‐wide SNP detection in the great titParus majorusing high throughput sequencing

Abstract

Identifying genes that underlie ecological traits will open exiting possibilities to study gene-environment interactions in shaping phenotypes and in measuring natural selection on genes. Evolutionary ecology has been pursuing these objectives for decades, but they come into reach now that next generation sequencing technologies have dramatically lowered the costs to obtain the genomic sequence information that is currently lacking for most ecologically important species. Here we describe how we generated over 2 billion basepairs of novel sequence information for an ecological model species, the great tit Parus major. We used over 16 million short sequence reads for the de novo assembly of a reference sequence consisting of 550 000 contigs, covering 2.5% of the genome of the great tit. This reference sequence was used as the scaffold for mapping of the sequence reads, which allowed for the detection of over 20 000 novel single nucleotide polymorphisms. Contigs harbouring 4272 of the single nucleotide polymorphisms could be mapped to a unique location on the recently sequenced zebra finch genome. Of all the great tit contigs, significantly more were mapped to the microchromosomes than to the intermediate and the macrochromosomes of the zebra finch, indicating a higher overall level of sequence conservation on the microchromosomes than on the other types of chromosomes. The large number of great tit contigs that can be aligned to the zebra finch genome shows that this genome provides a valuable framework for large scale genetics, e.g. QTL mapping or whole genome association studies, in passerines.

Genome-wide single nucleotide polymorphism (SNP) identification and characterization in a non-model organism, the African buffalo (Syncerus caffer), using next generation sequencing

“…In the African buffalo, Le Roex et al (2012) mapped 19-23% of their short reads (50 bp reads) to the domestic cow genome (Bos taurus). Our study confirms that using a genome of a closely related species as a reference standard provides a sufficient number of high confidence SNPs and offers a good alternative to characterize SNPs in non-model species without carrying out tedious steps of deep assembly of redundant contigs Van Bers et al, 2010).…”

“…Our average sequencing depth of the whole putative SNP database was estimated at 5x, which increased to 13x when estimated on the SNP set that passed the filtering criterion. This sequencing depth remains low compared to other studies (e.g., 25x (Van Bers et al, 2010), 58x ). The studies most similar to ours, to our knowledge, are that of Jonker et al (2012) (Branta leucopsis), which yielded an average sequence depth of 9.9x, and that of Le Roex et al (2012) (S.c. caffer), which yield an average sequence depth of 2.7x.…”

“…4). This decrease in predicted SNPs per position in Illumina sequencing reads was also reported in earlier studies Ramos et al, 2009;Van Bers et al, 2010), on which the current SNP detection pipeline is based.…”

“…Moreover, SNP predictions were tested by determining whether particular regions of sequence reads presented more SNP candidates than others. Previous studies have shown that tails of reads present excessively more sequencing errors, leading to false SNPs identification (Dohm et al, 2008;Van Bers et al, 2010). Within the SNP set passing all filtering criteria, no overrepresentation of SNP candidates in the ends of the reads was observed.…”

“…RRL approaches have been used to generate tens of thousands to millions of candidate SNPs with a genome-wide coverage for example in cattle (Tassell et al, 2008), turkey ) and great tit (Van Bers et al, 2010). Alternatively, SNP resources from one species could be used in a closely related species.…”