Bioinformatics Seminar
Time: 10:45am Tuesdays.
Venue:
Level 7 Seminar Room 2, WEHI1
17 July 2018
Comprehensive evaluation and characterisation of short read structural variant calling software
Daniel CameronWEHI Bioinformatics
In recent years, many software packages for identifying structural variants (SVs) using whole-genome sequencing data have been released. When a new method is published, it is commonly compared with those already available, but these comparisons tend to be selective and incomplete, and frequently show only modest improvements. The lack of comprehensive benchmarking across software tools presents challenges for users in selecting methods and for developers in understanding algorithm behaviours and limitations. Here, I present the results of a comprehensive evaluation and characterisation of 10 representative SV callers using typical human resequencing data, as well as simulations. I explain the impact on caller performance of event size and type, coverage, read length, library fragment size, and sequence context, and analyse the efficacy of ensemble calling and calibration of variant quality scores, concluding with some concrete recommendations for both software users and methods developers.