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

The use of this algorithm for genotype-to-phenotype models is just one of the potential 478 applications. Automated molecular assay design and degenerate-motif based phylogenetics are 479 examples of the downstream uses already being investigated. The ability to make use of the 480 . CC-BY-NC-ND 4.0 International license author/funder. It is made available under a The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.02.28.969782 doi: bioRxiv preprint latent data that is accumulating in databases, as well as novel surveillance data, is made more 481 tangible with this algorithm. Well-curated and richly annotated metadata promises to allow 482 machine learning and other data science techniques to unleash a torrent of discovery in genomics 483 at large. The mantra we are positing for the infectious and emergent diseases surveillance 484 community is "More data, Better data, Metadata." The techniques to unlock the potential of 485 data-driven genomic science are gathering momentum. The Vorpal algorithm for feature extraction was developed using the libraries and versions 560 delineated in the requirements.txt document located on the Github. The Vorpal feature 561 extraction algorithm has 3 steps, each corresponding to a script that becomes the Vorpal 562 workflow. 563