Venom metalloproteinase isoforms curated using genome alignments

The alignments of the collapsed isoforms from the Pacific Biosciences (PB) pipeline to the metalloproteinase (MP) gene complex (~1.2 mega base pairs and 30 annotated MP genes) were visually inspected using the Integrated Genome viewer (IGV) which identified many isoforms that aligned to different annotated MP genes but were incorrectly assigned (“collapsed”) to a single PB gene group. Additionally, many isoforms aligned to most of the coding exons for a gene but lacked sequence aligning to the start codon. Secreted venom proteins likely require a secretion signal so the isoforms lacking alignment to the start codon were removed from further analysis since they likely represent sequencing of degraded RNAs. These observations prompted us to refine the PB MP gene assignments by filtering the collapsed isoform set according to the following criteria: i) sequences that lack a start codon were removed ii) sequences that align within a gene interval bounded by 200 nt upstream of the start codon and 500 nt downstream of the stop codon were assigned to that gene. Filtering based on these criteria removed many sequences but retained a diverse set of putative isoforms from several classes (exon skipping, intron retention, truncation) (Dowell et al. Sup. Fig of MDC4). However, these curated sets of MP isoforms contained multiple full-length transcripts that were highly identical to each other. It is difficult to determine if these nearly identical transcripts represent biological differences (e.g. different untranslated region (UTR) lengths resulting from different transcriptional start sites (TSSs)) or artifacts of sequencing (e.g. degraded RNAs). Therefore, we specifically focused on accurate assessment of MPO1 expression with general attention on MP expression variation between individuals. To do this we tailored our approach towards understanding quantitative between specimen expression differences of reference genome-derived full length hypothetical MP transcripts.<br>