Steve Androulakis helps scientists and researchers get to grips with coding
'Should every scientist learn to code?' I was asked this in a recent programming workshop we ran for scientists.
It's a really good question. I was trained as a software engineer, but got a job in a biochemistry lab. If the cry for technologists in science appeared urgent to me then, it's kicking and screaming now.
'But why do we have to learn to code at all? Why aren't software engineers like you creating easy to use interfaces for us scientists so we can get on with our jobs?', another student asked me. This represents the other extreme, and is also a fair question. Should scientists really have to deeply learn such a foreign skill as programming to get by? How much should technologists like myself teach researchers to 'fish' on their own, and how much do we simply catch fish for them?
As manager of the Monash Bioinformatics Platform, this question is dear to me. We're a group of people that assist researchers in answering their data questions. We exist because data is suddenly beating down the door of an increasing number of medical and scientific research labs.
They have a scientific question, send their biological samples to a high throughput gene sequencer and receive a handful of incomprehensible files. Each file is tens of gigabytes in size.
This data requires complex arrangement using software before a scientist can interpret it. Arrangement of files this large can't be done on their own computers, so larger computers must enter the equation. The data processing software itself is as incomprehensible as the data and is command line based. This is often the point where they approach us.
In 2016, the vast majority of graduates who end up in a science lab were never educated in the skills required to make sense of data from modern scientific instruments. Groups such as our Bioinformatics Platform can help, but our group can't assist "We teach scientists to fish for themselves" everyone. This approach is not scalable nor sustainable.
As a result, we teach scientists to fish for themselves. Our group run a series of regular hands-on training workshops that are free to attend. They're a full day or two, and fill up very quickly.
Our workshops won't make scientists programmers or data analysts overnight, but they get a taste for what's involved. Most importantly, they now know we exist and that we're happy to assist them in working with their data on an ongoing basis.
Increasing complexity of data
Of course, sometimes we build infrastructure for the community. Several years ago, I created the MyTardis app under the guidance of a biochemist, Associate Professor Ashley Buckle. This was spurred by the very common desire of scientists to publish the raw, often hefty data behind their scientific results and cite it in their publications.
Last year at Monash, the eResearch team connected six different instrument facilities to MyTardis, thirty-three diverse instruments in total. For users of these electron microscopes, gene sequencers, mass spectrometers and other instruments, a little infrastructure and technology knowhow means their data is kept safe, and easily shareable.
The ever increasing amount and complexity of data in science will continue to pose challenges to researchers and technologists alike. The skills gap is closing, slowly. A combination of technologists teaching researchers to work with their own data and also providing them with technology solutions is what our group pride ourselves on. We hope this will help our scientific researchers withstand the growing avalanche of data into the future.
Steve Androulakis is the manager of the Monash Bioinformatics Platform at Monash University.
Photo courtesy of Philip Chan, Monash eResearch Centre