A Vision to Mitigate Bioinformatics Software Development Challenges

Abstract

Developers construct bioinformatics software to automate crucial analysis and research related to biological science. However, challenges while developing bioinformatics software can prohibit advancement in biological science research. Through a human-centric systematic analysis, we can identify challenges related to bioinformatics software development and envision future research directions. From our qualitative analysis with 221 Stack Overflow questions, we identify six categories of challenges: file operations, searching genetic entities, defect resolution, configuration management, sequence alignment, and translation of genetic information. To mitigate the identified challenges we envision three research directions that require synergies between bioinformatics and automated software engineering: (i) automated configuration recommendation using optimization algorithms, (ii) automated and comprehensive defect categorization, and (iii) intelligent task assistance with active and reinforcement learning.