Evolution of the Bovine TLR Gene Family and Member Associations with Mycobacterium avium Subspecies paratuberculosis Infection

Fisher, Colleen A.; Bhattarai, Eric K.; Osterstock, Jason B.; Dowd, Scot E.; Seabury, Paul M.; Vikram, Meenu; Whitlock, Robert H.; Schukken, Ynte H.; Schnabel, Robert D.; Taylor, Jeremy F.; Womack, James E.; Seabury, Christopher M.

SNP detection analyses for the resulting pyrosequencing data employed the Neighborhood Quality Standard algorithm [62, 63] implemented within CLC Genomics Workbench (v3.7.1), as previously described [29] . Putative SNPs were filtered using a method devised from a priori knowledge of biallelic controls (212 SNPs + 4 indels) [30] that were purposely seeded into the amplicon library. Briefly, we considered the possibility that some SNPs may only be found as one allele in a single elite sire (1/192 total alleles; see reference 30 for examples). Therefore, we filtered all putative SNPs predicted from our analysis of the pyrosequencing data using the following formula: 1/1926(Total SNP Cover-age) = Theoretical minimum number of reads, which represents the smallest number of reads required to shuttle putative SNPs into a validation workflow involving custom, allele-specific genotyping assays. This method proved valuable for the discovery and validation of many low frequency SNPs, including those that occurred as one allele for a single discovery panel sire (i.e., TLR5 putative nonsense SNP = 1/192 alleles in the discovery panel). For SNP discovery using standard dye-terminator sequencing reads, we used an alignment-based method of variant detection within the program Sequencher 4.6 [23, 25] . Briefly, high quality electropherograms were manually inspected for any evidence of a double peak. Individual nucleotide sites displaying any evidence of heterozygosity within$1 sequencing read were shuttled to our SNP validation workflow.