The path from bench to computational biology requires a piecemeal learning approach

The tagline of this newsletter is “Pathways to Computational Biology.” So far in 2024 I have interpreted this tagline to mean the pathway leading computational biology talent to solve problems in biotech/pharma. I have focused quite a lot on helping college/master’s/PhD students with some degree of quantitative training to walk this path and have achieved some success on that front. I’m pretty happy about that, but that’s just one very specific pathway. There are others that I have neglected thus far.

A pathway I have not paid enough attention to is the one that helps biologists without strong quantitative foundations to transition from bench work to compbio work. I’m talking about you wonderful folks who are performing cell/tissue culture, flow cytometry, and animal model experiments, just to name a few.

Bench biologists have great potential to become computational biologists. You already have intimate knowledge of the biology of the systems you are investigating. Many of you already think about your data quantitatively and have the desire to cement that thinking with more rigorous statistical analysis. You are actually latent computational biologists who just need a bit of help getting activated. But it’s also just hard to find the time to sit down to learn to code. Because you have to run experiments. Read papers. Plan the next experiment. So your learning has to happen piecemeal. I know, because I was in your shoes not that long ago.

I remember regularly having to run a 7-hour wet-lab protocol that included many 15-30 minute incubation steps. So during those short, built-in breaks in the protocol, I would stuff my tissue slides in the oven, set a timer, and rush right back to my computer to learn or practice data analysis in R. Then the timer would go off and I would have to immediately switch off my coding brain and switch on my tissue-handling brain. The context-switching was exhausting. I ran out of motivation to continue learning many times.

What would have helped me back then? A structured analysis that quickly led to an interpretable result to sustain my motivation. Imagine a short analysis script, light on theory, that in a few hours could take me from a raw dataset to a volcano plot for differential gene expression analysis. Seeing interpretable results so quickly would have given me the motivation I needed to dig into how the analysis was done. I would have wanted to practice the analysis again and again, each time gaining a deeper understanding of all the nuances that go into performing that analysis, each time becoming a better computational thinker. That is how I could have learned more effectively, and I think many bench biologists feel the same.

But just to be clear, I don’t think that structured content bypassing theory and leading to quick results is enough to get biologists fully proficient in doing compbio work. You need a lot more independent, intentional practice to get there. But it can at least generate enough momentum for new learners to keep going. It can help you “catch the bug” to do more compbio work. And for many bench biologists, that might just be what they need to find their path to compbio work.

All this is to say that I am now working on that kind of structured content. This is still in line with the learning philosophy promoted by Figure One Lab, except it will be tailored more to bench biologists.

Stay tuned.