Simulation and Statistical Learning Methods Link Aflatoxin B1 Induced High-Resolution Mutational Spectra in Mice with Human Hepatocarcinogenesis

Human hepatocellular carcinoma (HCC), the second leading cause of cancer deaths worldwide, is a complex genetic disease with numerous etiologies. HCC incidence is highest in sub-Saharan Africa and Southeast Asia due to chronic exposure to the naturally occurring mutagen aflatoxin B1 (AFB₁) and to a high endemic prevalence of hepatitis B and C viruses which predispose the liver to cirrhosis and HCC development. In this work, we investigate the risk factor AFB₁ by examining its ability to cause base substitutions in all 96 canonical trinucleotide contexts. A Monte Carlo method was built to model liver tissue exposed to mutagens of varying specificity in order to identify the number of mutations needed in an experimental regime to confidently represent a biological effect in downstream analysis. To characterize the mutagenic pattern of AFB₁, four gpt delta transgenic mice were dosed with AFB₁ at four days of age and then sacrificed at ten weeks. Duplex Consensus Sequencing (DCS) was used to interrogate the ultra-rare variants found in a 6382 nt transgene fragment harvested from the liver. A DCS analytical pipeline was then built to assemble an AFB₁ exposure spectrum and compare it to the mutational spectra extracted from 314 sequenced human HCC tumors using statistical learning methods.