Allowing for non-binomial sampling of alleles in sequencing-based genotypes

Outline

• Analysis of Genotyping By Sequencing (GBS) data
• How can we detect insufficient deduplication without “gold-standard” inbreeding estimates?
• Can we make adjustments later in the analysis process (without flowcell position information)?

Conclusions

• Deduplication is a useful tool with patterned flowcells
• But parameter settings unclear
• Using non-binomial allele sampling models can correct remaining examp duplication effects
• Suitable training sets needed to parametrize models
• Assaying libraries across 2 lanes of a flowcell provides suitable data