NCMM's strong presence was marked by the expertise of our early career researchers from NCMM’s Marieke Kuijjer’s and Anthony Mathelier’s research groups. Katalin Ferenc and Ieva Rauluseviciute, both PhD students in bioinformatics in the Mathelier group, and Ladislav Hovan, a postdoctoral fellow in the Kuijjer group, shared their knowledge and skills in a workshop attended by a 30-strong cohort. The workshop focused on improving software quality in bioinformatics through testing, a crucial aspect of the field.
The workshop described systematic, automated, and reusable software testing methods, and why it matters when developing bioinformatics software in particular. The facilitators led the discussion among the participants about the difficulties of thinking of appropriate tests. They highlighted that thinking about good tests for scientific software can be thought of as a tool, which helps to ensure the robustness of any data analysis software. The team demonstrated the advantages of code testing, such as improving the functional stability and maintainability of a particular software providing code examples, and by encouraging participants to test their code in groups. They also discussed the collective approach to software development and code testing in bioinformatics undertaken between the Mathelier and Kuijjer groups to improve coding culture by using three overarching pillars: software quality seminars, code reviews, and resource sharing.
Feedback from one participant showed how important this type of workshop is. They expressed their “gratitude for the most valuable part of the conference, with the advice provided on testing and general tips on writing code better being incredibly helpful.” They also shared that they “spent the entire afternoon and evening going through my code and improving it (and I'm far from done with that). The workshop has inspired and motivated me a lot, and it will change the way I write code in the future.
Finally, Ladislav Hovan gave a presentation on “Spatially resolved gene regulatory networks”, introducing the new STOAT (Spatial TranscriptOmics to Assess Transcriptional regulation) approach to deal with sparse gene expression in Visium spatial transcriptomic data, in order to create gene regulatory networks. STOAT was validated on breast cancer and melanoma datasets. Anthony Mathelier, who was part of the program committee, also chaired a session.