I believe there are two cultures of bioinformaticians in academia: those that build tools and those that combine tools. The tool-builders enable analyses, the tool-combiners carry out analyses. Both contribute to the betterment of the world, but only one has a future in academia.

The builders

Tool builders are those bioinformaticians who come more from a computer science background, who delight in strange data-structures and build complex code-bases to enable analyses. They (usually) do not particularly care about biological minutiae, they want to build software. They write papers like ‘A novel Trie-like data structure for DNA k-mers’ or ‘Painting basespace in colourspace’. They know a lot about threads, complex data structures customised to biological data, and have seen every bug under the sun. Their God is Heng Li and yes, they have built a short-read aligner. Builders have a comfortable career in academia.

The combiners

Tool combiners are those bioinformaticians who come more from a biology/wet-lab background. They have biological questions that need answering, their tool of choice happens to be a computer. Some, like me, specialise out of the lab and onto the couch with a laptop. These people do not care for technical minutiae, they want to solve problems and answer biological questions. In a regular analysis, combiners use eight tools they way they are intended to be used and do some violence to two more tools to arrive at an answer. Combiners know a lot about the weirdnesses of their species of interest, they know which tools work and which don’t. They usually solve problems for others, so they usually end up somewhere in the middle of the paper’s author-list. Combiners do not have a comfortable career in academia.

A quick note

Of course, every bioinformatician sits somewhere between these two cultures. Cultures are semipermeable like membranes, they are not hard like human-made borders.
I wonder where data-scientists see themselves within these two cultures. My hunch: more on the side of the combiners.

On leadership in academia

Much has been written about the strange ways academia evaluates itself, and all the negative repercussions of those ways from mental health crises to risk aversion. My feeling is that the main reason is a strange fixation on a specific version of leadership. Academia’s idea of leadership has its roots in the idea of the lone genius scientist, the version of scientists that is mirrored in the Nobel Prize’s fixation on between one to three researchers per Prize. It is heavily implied that these achievements were made possible by a few lone geniuses, not by huge (underpaid, overworked) teams of PhD students, postdocs, research assistants, lab managers, and all the other unsung heroes of science.

This fixation on lone leaders leads to the strange way scientists are evaluated: you have to be a leader in everything. Every grant application asks you how you’re the leader in your field. Every university evaluates its academics based on the number of papers you’ve been first author on (therefore, have led), and how many grants you’ve been a lead investigator on. These two numbers are the only thing that counts to academic evaluation. Much has been written how detrimental that is to science, and yeah it’s a tragedy, but not one that anyone shows any effort to change.

On leadership in bioinformatics

How does this fixation on leadership influence our two supposed cultures? Leadership-fixation favours the builders and pushes the combiners out. The builders, by the nature of being tool-builders, can show the powers that be that they’re leaders. Builders build software, write software papers, and get grants to write software. The way that builders work maps well to the way academia evaluates: builders have many first-author papers and lead on software-heavy grants. I don’t think builders ever planned for it to happen this way, it’s just chance that their field and academia’s goals aligned.

Combiners, on the other hand, are the unlucky ones. If they’re not careful they’re the perpetual middle-author on many, many papers. These papers may end up being highly cited or highly influential, but that won’t matter to the evaluators, as the combiners are not leading on those papers. Combiners obviously run their own research, but they’re often dependent on others to collect the samples or run the experiments, so the danger of getting ‘pushed into the middle’ is high. (Later edit: This does not mean that builders, in the age of constant budget cuts and low research funding, have it easy in academia. They have to struggle to fight for the same dwindling pot of funds.)

So what?

Combiners are needed in academia, every lab is drowning in data that they cannot analyse. I know several lab leads here that need combiners but cannot find any to hire. I know several bioinformatics support units at universities that got shut down even though they were at 100% capacity: to their respective leaderships, they did not bring in enough funding and did not lead on enough papers (how???). I’ve left academia, and the best ‘combiners’ I know are now data scientists in fintech, government, or health companies. Their combined skills are lost to academia. It’s not like labs just stop generating data, on the contrary, the volumes get bigger and bigger.

Where to next?

As written in countless other pieces, academia needs to change how it evaluates itself, away from ’lone leaders’ and towards group efforts. That needs a cultural change, not just in academia, but also in how wider society sees scientists: not as lone geniuses with white hair sitting in their ivory towers, but regular people with mortgages and children, working together.

I’m too old to wait for that change to happen. I’m already half-way done with my career: science progresses by funerals.

P.S.: The Research Engineer pathway is a partial answer to the question of career. ResEngs are not academics, but academic staff. But then you have a new problem: university career pathways are generally managerial in nature, there are no engineer-heavy pathways like FAANG and other engineering companies have. At some point ResEngs will run out of career and additional earnings.