Analytical Validation of a Computational Method for Pharmacogenetic Genotyping from Clinical Whole Exome Sequencing.

Germline whole exome sequencing from molecular tumor boards has the potential to be repurposed to support clinical pharmacogenomics. However, accurately calling pharmacogenomics-relevant genotypes from exome sequencing data remains challenging. Accordingly, this study assessed the analytical validity of the computational tool, Aldy, in calling pharmacogenomics-relevant genotypes from exome sequencing data for 13 major pharmacogenes. Germline DNA from whole blood was obtained for 164 subjects seen at an institutional molecular solid tumor board. All subjects had whole exome sequencing from Ashion Analytics and panel-based genotyping from an institutional pharmacogenomics laboratory. Aldy version 3.3 was operationalized on the LifeOmic Precision Health Cloud with copy number fixed to two copies per gene. Aldy results were compared with those from genotyping for 56 star allele-defining variants within CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP3A4, CYP3A5, CYP4F2, DPYD, G6PD, NUDT15, SLCO1B1, and TPMT. Read depth was >100× for all variants except CYP3A4∗22. For 75 subjects in the validation cohort, all 3393 Aldy variant calls were concordant with genotyping. Aldy calls for 736 diplotypes containing alleles assessed by both platforms were also concordant. Aldy identified additional star alleles not covered by targeted genotyping for 139 diplotypes. Aldy accurately called variants and diplotypes for 13 major pharmacogenes, except for CYP2D6 variants involving copy number variations, thus allowing repurposing of whole exome sequencing to support clinical pharmacogenomics.

Potential Utility of Pre-Emptive Germline Pharmacogenetics in Breast Cancer.

Patients with breast cancer often receive many drugs to manage the cancer, side effects associated with cancer treatment, and co-morbidities (i.e., polypharmacy). Drug-drug and drug-gene interactions contribute to the risk of adverse events (AEs), which could lead to non-adherence and reduced efficacy. Here we investigated several well-characterized inherited (germline) pharmacogenetic (PGx) targets in 225 patients with breast cancer. All relevant clinical, pharmaceutical, and PGx diplotype data were aggregated into a single unifying informatics platform to enable an exploratory analysis of the cohort and to evaluate pharmacy ordering patterns. Of the drugs recorded, there were 38 for which high levels of evidence for clinical actionability with PGx was available from the US FDA and/or the Clinical Pharmacogenetics Implementation Consortium (CPIC). These data were associated with 10 pharmacogenes: DPYD, CYP2C9, CYP2C19, CYP2D6, CYP3A5, CYP4F2, G6PD, MT-RNR1, SLCO1B1, and VKORC1. All patients were taking at least one of the 38 drugs and had inherited at least one actionable PGx variant that would have informed prescribing decisions if this information had been available pre-emptively. The non-cancer drugs with PGx implications that were common (prescribed to at least one-third of patients) included anti-depressants, anti-infectives, non-steroidal anti-inflammatory drugs, opioids, and proton pump inhibitors. Based on these results, we conclude that pre-emptive PGx testing may benefit patients with breast cancer by informing drug and dose selection to maximize efficacy and minimize AEs.

Translating cancer genomics into precision medicine with artificial intelligence: applications, challenges and future perspectives.

In the field of cancer genomics, the broad availability of genetic information offered by next-generation sequencing technologies and rapid growth in biomedical publication has led to the advent of the big-data era. Integration of artificial intelligence (AI) approaches such as machine learning, deep learning, and natural language processing (NLP) to tackle the challenges of scalability and high dimensionality of data and to transform big data into clinically actionable knowledge is expanding and becoming the foundation of precision medicine. In this paper, we review the current status and future directions of AI application in cancer genomics within the context of workflows to integrate genomic analysis for precision cancer care. The existing solutions of AI and their limitations in cancer genetic testing and diagnostics such as variant calling and interpretation are critically analyzed. Publicly available tools or algorithms for key NLP technologies in the literature mining for evidence-based clinical recommendations are reviewed and compared. In addition, the present paper highlights the challenges to AI adoption in digital healthcare with regard to data requirements, algorithmic transparency, reproducibility, and real-world assessment, and discusses the importance of preparing patients and physicians for modern digitized healthcare. We believe that AI will remain the main driver to healthcare transformation toward precision medicine, yet the unprecedented challenges posed should be addressed to ensure safety and beneficial impact to healthcare.