Microsatellite Recovery
Microsatellite genotypes were recovered at a higher rate than expected based on failed results from low quality samples in PCR. The LQMS recovered an overall PCR success rate of 21.4% (calculated from the average success for all recorded PCRs of the LQMS from Appendix 1) yet recovered 42% genotyping success (calculated from a total of 30 recovered genotypes from the possible 72 for the LQMS, Table 4). This percentage did not include removal of problematic genotypes. When only genotypes that were agreeable were included, this reduced the 30 genotypes to 6 or 8.3%. Despite this very low rate of confirmation among the LQMS this study quantifies rates for genotyping success via GBS on poorly amplifying museum specimens for the first time. Alternatively, for HQMS the rate of recovered genotypes was 91.7% (66 of 72), and 86.1% (62 of 72) of agreeable genotypes. For the tissue sample 100% (16 of 16) of the replicates resulted in a genotype, with 87.5% (14 of 16) confirmed by the second genotype recovered. This provides robust support that the rate of disagreement shown in the HQMS is negligible and only 1.4% less than the tissue sample.
Samples which routinely amplified produced reliable genotypes. Low quality samples had variable genotypes which seemed stochastic and unreliable across various PCR replicates. Additionally, it appeared that these variable genotypes may have resulted from differences between replicated PCR. This was apparent when sequencing reads were bioinformatically combined as well as from the pooled run. Short microsatellites amplified more frequently and certain microsatellites had more reliable calls than others. This study has shown that samples which do not reliably amplify as assessed by gel electrophoresis may be prone to poor performance in PCR and result in inaccurate genotypes.
We were not able to compare the resulting genotypes recovered here to other population genetics studies on G. oregonensis , as this is the first time such a study has occurred in this species. The study performed by Barbian et al., (2018) used GBS to re-genotype chimpanzees with known life-history data and previous capillary electrophoresis (CE) genotypes, and noted a shift of 1-3 bp in genotype results between the CE and GBS data. They also noted the recovery of additional allelic diversity, which was traditionally lost in homoplasy (and confirmed by pedigree analysis). We also recovered homoplastic events (detailed in Appendix 5) in our data. Darby et al., (2016) noted a 44% increase in alleles due to homoplasy in their dataset as well, and increased alleles from 164 to 294. In other words, 130 novel alleles were recovered from GBS over CE.