The BSI is delighted to have a longstanding collaboration with John Cole (Glasgow Bioinformatic Core), to equip wet-lab immunologists, clinical scientists and other life scientists with the skills and confidence to perform their own bioinformatic data analysis. The course has had over 3,000 attendees since 2019, with a mean rating of 9.5/10 for content and delivery. We caught up with John to discuss what makes these courses so popular and what the future holds.
Tell us a little about yourself; when did you start working with bioinformatics?
I’ve had quite an unusual career: I left school at 16 and trained to be a landscape gardener, and then studied genetics in Glasgow, but didn't like the wet lab so I then became a chef for a while. I realised that I love science but don't like the wet lab, so I went into bioinformatics and did a master’s degree about 16 years ago. I went on to work in the field of cancer research, so I did quite a lot to do with ageing and cellular senescence and eventually moved into immunology. Eventually, I moved on to managing a small bioinformatics unit and training others.
Something I noticed very early on was that there are a lot of misconceptions and fear about learning how to do bioinformatics, but also how important it is for people from a range of research areas to learn it. Because even though I was a good bioinformatician, I couldn't do everything – I didn't know a lot of the biology and didn't have time to learn it.
Wet-lab immunologists often don’t have a coding background and view bioinformatics as quite daunting. How do you tailor your training to support those people?
The important thing is to understand where people are coming from. I didn't learn to code until I was nearly 30. So, I can remember that feeling of starting something new that is very different. Coding, like any skill, you can learn. It is the application of logic and that's what scientists do all day long.
What’s key is making well-structured, well thought through content that is consistent in how it’s delivered. It's always a lecture followed by a tutorial. The lecture is not there for anything other than context around what the code is supposed to do and then you code it yourself with the demonstrators there. Any lesson that’s important is repeated, but from a different angle, so you’re seeing the same thing from lots of different viewpoints and start to build up a working understanding: how R works, what you're trying to achieve and how you do it, and how to make beautiful plots.
The main thing to remember is that it takes three or four days to get over the hump. That's why people often end up using Excel, Prism or GraphPad, because you can do it immediately. With coding, you have to put a solid few days in until you really start to see the effects, and maybe a full week until you're making beautiful plots. What we try and do is lower the difficulty in getting to that point by keeping it friendly, by taking it slow, by explaining everything in great detail; what every little bracket does, why you're doing it, and how to think about it.
What we do differently from other bioinformatics courses is that at the beginning we demonstrate why it's worth learning. Another thing we do differently is that we don’t teach chronologically. The coding you need to know to do processing is the most difficult to learn. We start at the analysis, that’s when you're doing actual biology and making plots. It's very easy to engage.
How many people would you say you have trained at this point?
Over 3,000 PhD students, postdocs and PIs, and I think for masters students and undergraduates, it's probably another 2,000. It's getting quite a big number now!
Big data approaches are becoming more prevalent across biology. Where do you see the biggest skill gap in immunologists?
I guess the obvious answer is finding people that are experienced to do the analysis. It is less of a problem than it was because people do seem to be getting trained. Glasgow is particularly ahead of the curve in terms of training undergraduates and master’s students, but other universities maybe aren't quite there yet.
There's also a big gap in terms of experienced informaticians. There are now quite a lot of people that have six months to a year. What we don't have is many people who are very experienced. Having someone to talk to that knows what they're doing is also very important. That's something we offer as part of the courses, too. We run drop-ins once a week for most of the year, and you can come and talk to me or one of my colleagues who know what we're doing.
At what career stage would you encourage people to attend bioinformatics training?
It's open to any career stage. I think the main thing is to learn bioinformatics before you need it. If you're planning an omic project or if you think that's going to be your next postdoc or that's going to be your first PhD or that's what you're putting into your next grant, then you should learn it, right? Even if you decide that coding is not for you and it's for people in your lab, it's still very useful to understand. You will come across so many things in papers, and have to review them, and it’s useful to be familiar with these base concepts.
Does learning bioinformatics and the concepts behind it help with career prospects for those who don't stay in academia?
Absolutely. Beyond wet-lab research, industry still needs bioinformatics expertise. If you're moving into industry, or even sales you might need to sell something which needs that to some degree; it's useful. I think people that learn coding can then go on to other things. When I studied bioinformatics, it was a one-year master’s, and half the class went off to become programmers. One of the people went off to work for an investment bank. So, there's lots of opportunity that it opens up. It's not a huge investment where you might discover something that you really enjoy in its own right. So, I wouldn't say that bioinformatics is relevant outside medical science; but coding is generally of use.
In terms of the people that have attended your training and the feedback you've received, how have you shaped and evolved the courses as time has gone by?
We've run the course more than 50 times now and used feedback to hone it as time has gone on. Certainly, in the beginning, there were questions like, should we run it for one-week full-time or should we run it for two weeks of mornings? What works best for a scientist?
We’ve also reviewed which bits are not explained so well, and refined what was not clear in the tutorial book. What we also do with each of the courses, is have a kind of bonus section. So, if there's something that comes up often, for example, producing a rock curve then we provide bonus code and a tutorial. So, we've tried to do that for other things that we can't fit into the main thread of the course.
And this year we will be offering a self-paced learning option in addition to the live formats that already exist. How are you feeling about that?
It's exciting to offer a range of things for a range of different people. People think in different ways, so it's exciting to try and accommodate as many different people as we can. There are also plenty of people that can't commit to two weeks of mornings, especially PIs, clinical people, and so on; that's difficult. We find that when we’ve run courses for groups of people self-paced only, most of them just get on with it and do it no problem whatsoever.
What also is very good is if people sign up together in the same group, then they can talk to each other. Being part of a group chatting to each other and empowering each other is a great way to learn. And there will always be the drop-ins, so people can come and speak to us. We set up a relaxed, camaraderie style place where you can talk to us. So, if it's not working, then you just have to talk to us to get over the hump, until you get onto something you find easier.
Interviewed by Laura Cox
