Friday 10 August 2018

Two half ideas is a solution

It's very quiet at The Institute: all the lecturers are away on their 2 month vacation. Any graduate students under their supervision are no long supervised and (therefore) not present. In my day, I worked from dawn to dusk in Summer and from full dark to full dark in Winter. There was always stuff to be done even if the Boss was distracted, which was often because he many interests. Last Thursday, I was invited to a session with two of our Effectives - Eff.I and Eff.II - who were on campus because they also realise that stuff has to be done or the project will stagnate, they personally will learn nothing and their years in further-further education will be wasted.

One of the research groups has sequenced the genomes of a couple of soil bacteria which seem to have something interesting, and with  commercial potential, in their biochemical toolbox. It doesn't matter what it is: the bugs could sequester heavy metals like cadmium or convert ethanol into single malt whiskey or make a fabulous new antibiotic. Eff.I has been tasked with searching the chunks of DNA which have come back from the sequencing company for bits which are similar to known enzymes involved the biochemical pathway of interest. He's learning to program in Python because that is the majority language for bioinformatics this year. In my day it was Perl. You cannot learn programming as a disembodied hobby: you must have a task that needs doing which can only be carried out by coding. It's also really difficult to learn how to write code all on your own. When I was going for my PhD, I wrote Fortran to calculate and print genetic distance GD matrices. I was the only coder in the Biology Department, so I had nobody to help, it was hard graft and not much fun and took me a long time. But it sure beat calculating GDs with a pocket calculator because the matrix grew to include 35 populations and 35 * (35-1) = 1190 separate pairwise GDs to calculate. It was a problem that 'scaled up' if you could code it. Luckily Eff.I has met Eff.II [thanks to matchmaker me] who did a research project with me last year and is now doing work experience in UCD. Because that internship is mostly coding, he spends most days 'working from home' also learning Python, round the corner from The Institute.

We met in a quiet part of the concourse at one of those tables with one leg shorter than the others. Eff.I showed us his code <whooof - over the head> and started to talk jargon which Eff.II was making a better fist than me of understanding. We were on firmer ground when Eff.I showed us some graphs which his code had generated: they weren't pretty and were even a little bit confusing but we could see the results and could then discuss how they could be made more fit-for-purpose. That purpose being to show Eff.I's computer-illiterate bosses that he was getting down with some of the data. Eff.II was able to share some fragments of his code that might produce a better, more informative picture.

While they were talking shop, I got up and rootled through the nearest wastebin for a bit of cardboard to wedge under the wonky table leg. I may be quite long-in-the-tooth and out of practice with the software but hardware I can bodge with the best of them. Viva la bricolage!

No comments:

Post a Comment