Application of machine learning to large in-vitro databases to identify cancer cell characteristics: telomerase reverse transcriptase (TERT) expression

Advances in biotechnology and machine learning have created an enhanced environment for unearthing and exploiting previously unrecognized relationships between genomic and epigenetic data with potential therapeutic implications. We applied advanced algorithms to data from the Cancer Dependency Map to uncover increasingly complex relationships. Specifically, we investigate characteristics of tumor cell lines with varying levels of telomerase reverse transcriptase (TERT) expression in liver cancer. The findings indicate that the effect of CRISPR knockout of Histone Deacetylase 1 (HDAC1) and numerous individual respiratory complex I genes is strongly related to the level of TERT expression, with knockout being particularly efficacious at killing or inhibiting growth of tumor cells with low levels of TERT expression for HDAC1 and high levels for Complex I genes. These findings suggest key biomarkers for therapeutic efficacy and yield novel potential pathways for drug development and provide further proof of principle for the potential of artificial intelligence in oncology.

Automated summary

Advancements in biotechnology and machine learning have created a conducive environment for exploring relationships between genomic and epigenetic data with potential therapeutic implications. The study found that CRISPR knockout has strong efficacy in inhibiting growth of tumor cells with varying levels of TERT expression, providing key biomarkers for therapeutic efficacy and potential pathways for drug development. These findings further support the potential of artificial intelligence in oncology.