Vorpal: A novel RNA virus feature-extraction algorithm demonstrated through interpretable genotype-to-phenotype linear models

Phillip Davis; John Bagnoli; David Yarmosh; Alan Shteyman; Lance Presser; Sharon Altmann; Shelton Bradrick; Joseph A. Russell

isolates. The conservation of the CTA leucine codon in the human-isolate predictor is 227 noteworthy because this codon is one of the rare leucine codons in the human genome, with a 228 relative frequency of 7%. Alternatively, the TTA codon being more predictive for swine isolates 229 is notable because while TTA also only has a 7% relative abundance in humans, its abundance in 230 pigs is 6% while the CTA codon is less rare (13% relative abundance) 19 . This mammalian 231 adaptation separates it almost entirely from any avian examples and there appears to be a fitness 232 gradient. When it appears in mammals, there is a higher incidence of the uncommon leucine 233 codon at this location. Table 4B , and this relationship between motif 242 mapping frequency following a codon rarity gradient in mammals, and the inverse in birds, is 243 visualized in Figure 4 . that the TTA codon for leucine plays in increasing the probability of a Swine isolate classification 249