Bioinformatics: Benchmarking Variant Callers & Analysis Tools

Next generation sequencing (NGS) has allowed for the quick turn around time of whole genome and transcriptome sequencing of cancer patients, which enables a detailed look at somatic changes in tumours.

However, there are many confounding factors in the identification in mutations, structural changes and gene expression changes, such as systematic bias, sequencing error and varying tumour content from the tumour sample. Although there are a large number of tools available for analyzing somatic variation, choosing the right tool can be very difficult.

Here we present several examples of the testing and benchmarking of tools that have been done at the Genome Sciences Centre, and the process that we used to select a tool for our analysis pipelines. One common limitation of benchmarking exercises is the lack of ground truth. We will also present a somatic variant data set which we constructed based on the sequencing of a metastatic melanoma cell line, COLO-829 and its matched normal.

With these results, we were able to generate a somatic reference standard for cancer genome sequencing and benchmark a number of somatic variant callers.

To view the slides from this presentation, visit:
https://www.westgrid.ca/events/bioinf...

For more information on other WestGrid training events, visit:
https://www.westgrid.ca/events