Experiences of Testing Bioinformatics Programs for Detecting Subtle Faults

One of the biggest challenges for conducting automated systematic testing on scientific programs is the oracle problem. This challenge is especially prevalent in the field of bioinformatics due to the inherent complexity of these programs. In this paper, we explore two approaches: pseudo-oracles and metamorphic testing for conducting automated systematic testing of bioinformatics programs. We use BBMap: an open source genome alignment tool as the system under test to evaluate the effectiveness of the above two approaches for identifying subtle faults.