In-house or out-house: Is it time to outsource bioinformatics?

This post is based on two recent blog posts on the state of the bioinformatics industry. In them, the authors (Philipp Zentner and Tatsiana Aneichyck) point out the issues currently plaguing the field of bioinformatics and the role that the private sector could play in addressing them.

Bioinformatics is still the preserve of academia (although the situation is changing at least in the area of next generation sequencing), and as such subject to its rules. Two aspects of academic life have in particular a profound impact: firstly, the short-term nature of positions at the most junior level (PhD students and Postdocs), where most of the actual software development and bioinformatic analysis takes and secondly, the fact that publication of novel algorithms and tools is rewarded, but not their maintenance. As a consequence, established software quickly becomes obsolete and new tools and standards keep arising, forcing researchers to constantly adapt their pipelines accordingly and slowing down analysis.

The collaborative nature of academia, which is essential to both push research further but also needed by researchers to apply for grants to continue their work, also brings its own complications. The increasing popularity of genome and transcriptome sequencing, for example, means the bioinformatics units have to spend more of their time analysing data from such collaborative projects, rather than developing new tools and algorithms. And since their services are essentially free to academic collaborators, demand is continuously increasing and often offloaded to already overworked and inexperienced PhD students and young postdocs. Analysis quality can suffer as a consequence, a fact that has not passed unnoticed in academia itself. This is also consistent with the recent reports on unreliable data analysis in scientific articles (1-2).

It would make sense for academic teams to outsource their analysis to  bioinformaticians offering their paid services, either as freelance actors or as part of a for-profit company. They provide both expertise and also have a strong incentive to maintain and update their pipelines over the long term, resulting in more efficient and quicker analysis.

This is even more crucial as bioinformatics is becoming more complex and specialised. It is increasingly hard for a bioinformatician (let alone a student) to become an expert in all the areas and subareas of bioinformatic analysis. And yet, in academia bioinformaticians are often expected to deal with vastly different data types analysis (e.g. protein structure, phylogenetic tree analysis and epigenomics) at a deep level. Thus, outsourcing tasks for which there is no competence in a group saves time and also ensures the analysis is performed by a bioinformatician who is highly familiar with the data and the methods of analysis.

This paradigm change in bioinformatics will bring with it new challenges and opportunities, both within academia, with a renewed focus on creative research and for start-ups focusing on data analysis.

References

(1) Smaldino PE, McElreath R. The natural selection of bad science. R Soc Open Sci. 2016;3(9):160384. Published 2016 Sep 21. doi:10.1098/rsos.160384.
(2) Brown AW, Kaiser KA, Allison DB. Issues with data and analyses: Errors, underlying themes, and potential solutions. Proc Natl Acad Sci U S A. 2018;115(11):2563-2570. doi:10.1073/pnas.1708279115.