Investigation of bioinformatic pipelines for variant calling using RNA-seq data

Sammanfattning

This project has investigated tools that could be used for somatic (non-inherited) variant calling using RNA-sequence (RNA-seq) data from patients with non-small cell lung cancer (NSCLC). Identifying specific genetic variation causing growth of cancerous cells have been a major challenge within cancer research and diagnostics. Variant calling uses sequenced data to analyze differences in the genome for further usage in targetable treatments. The aim of this thesis was to propose two variant calling pipelines for variant calling using RNA-seq data and test the performance of these. Variations investigated were single nucleotide variants (SNVs) and insertions and deletion (indels). This has mainly been done on DNA-seq data but research has shown that RNA-seq can provide a faster and more cost effective alternative. Additionally it has the potential to uncover novel mutations in well-known biological pathways and better detect low frequency variants. A comprehensive literature study was carried out to decide upon two variant calling tools for testing where Mutect2 and VarDict were chosen. The testing was performed on 197 patients with NSCLC with a focus of finding variations in the genes KRAS, EGFR, BRAF, ERBB2 and MET that caused protein changes. The result showed a possibility of finding clinically important mutations using RNA-seq data where the callers performed fairly similar but Mutect2 slightly better. No significant differences between the callers could be seen regarding capturing SNVs or indels. This study was carried out on a relatively small data set and could be further investigated on larger data sets to validate the result.