GATACA is developing novel software solutions for virologists. Funded by Small Business Innovation Research (SBIR) awards from the National Science Foundation (NSF) and the National Institutes of Health (NIH/NIAID), our current focus is on hepatitis B virus (HBV) bioinformatics aided by deep machine learning (ML). Our algorithm, GAT is a novel framework for de novo assembly of HBV NGS output into quasispecies that we are now empowering with artificial intelligence (AI)—specifically ML methods to learn the ‘syntax’ of the infecting strains. GAT/ML™ trained on HBV variation using our proprietary database benchmarked as superior against leading viral assemblers in simulation studies and tests on patient samples, and—is currently undergoing further optimization for enhanced analytics of patient samples. Our abstract on a soon-to-be published patient case-study demonstrating GAT/ML’s utility to identify rare variants and quasispecies will be presented at EASL (London, August 2020¹).

Other development underway includes statistical modeling of quasispecies evolution, and prediction models to forecast disease progression.

GAT/ML™ trained on a vast knowledge base of HBV variations processes a high NGS volume and profiles existing dangerous HBV quasispecies with speed, precision and unprecedented reproducibility.

Analytics for HBV and HCV are ongoing in-house. GAT/ML™ v1.0 for HBV will be available for testing in Summer, 2020.

Using our proprietary first ML module trained on HBV variation, the sequences clustered into distinct major genotypes and subtypes, focused GAT assemblies and improved haplotype accuracy

• • On datasets with ≥1,000,000 and as few as 1,000 reads
  • Quantifies the fraction of each subtype present in the dataset