The curriculum task force of the International Society for Computational Biology’s (ISCB) education committee recently published a paper in Plos Computational Biology that describes their efforts to come up with guidelines to design bioinformatics courses and curricula that equip graduates with the skills they need to work in a variety of settings including core facilities, biomedical research laboratories, pharmaceutical companies.
The paper describes the results of several surveys which the task force conducted to get a better sense of the skills and abilities that aspiring bioinformaticians need to have and includes a list of “core competencies” that they believe, based on their analyses, are needed by three different types of bioinformatics professionals. They hope that educators and trainers use these core competencies as guidelines when making decisions about what courses to include in their programs.
This week, BioInform spoke with Lonnie Welch, a professor of electrical engineering and computer science at Ohio University, and one of the members of the curriculum task force about the paper and their continued efforts to come up with general guidelines for bioinformatics education. What follows is an edited version of the conversation.
In 2012, the ISCB committee asked for the community’s input on a draft consensus curriculum for bioinformatics. What were some of the responses you got?
This report was published in Plos Computational Biology in 2012. Basically we surveyed the ISCB education committee and we surveyed the EMBnet and we got about 40 responses representing 20 different countries. The main thing that came out of that survey was a set of topics for an initial working curriculum. There were five topics in the area of computing, math, and statistics, and, interestingly, there were also five topics in the area of biology and chemistry.
Also in 2012, you established three working groups to focus on three specific areas of education. What were you hoping to accomplish?
We wanted to capture a broad perspective of what’s needed for success in bioinformatics and computational biology. So we put together three working groups. One of them was a group that surveyed core facility directors. That was led by Fran Lewitter. We got about 30 directors of core facilities who responded to that survey. I put together a group that surveyed career opportunities and they looked at 75 different job listings from the ISCB career pages. And the third group looked at existing educational programs and that was directed by Russell Schwartz. He surveyed more than 150 programs and he focused only on programs that either awarded degrees or certificates and that were exclusively called either bioinformatics or computational biology.
In one of the tables in your paper you list the requirements for success in bioinformatics and computational biology that you got from surveying core facility directors and from job opportunities. There’s a very wide variety of skills required and designing a curriculum that would comprehensively addresses those needs seems like an uphill task. How did you distill those down into something that’s more general?
Once we had the list of skills that emerged from the surveys, I went and looked at the accreditation criteria for computer science programs and they had a list of core competencies and I found that list to be a good starting point. We gradually morphed that list until we had something that brought into it the additional skills that were needed from biology and from the other related areas.
Did you come across any unexpected requirements in conducting the surveys or analyzing the job opportunities?
One of the interesting results was that the core facility directors were asked what skills are lacking in recent hires and also what skills are needed for success and it turns out that those two sets are similar to each other.
What specific skills were lacking?
They are really looking for somebody who has the ability to begin with the raw dataset and then to independently perform the complete computational analysis that’s needed to interpret that dataset.
But isn’t that the very essence of bioinformatics training? Why aren't students prepared for that?
It seems like it should be the essence of bioinformatics training. I think we also get focused on the details of bioinformatics methods. The details are very important but it’s also critical to understand the role that bioinformatics has within the complete scientific discovery process. I think that a lot of us could probably do a better job of teaching our students that complete end-to end-process.
The survey also showed that things like interpersonal skills, time and project management skills, and so on are requirements for success. Would it be worthwhile trying to incorporate classes focused on teaching some of those skills into a curriculum? So maybe adding communication classes for instance?
Communication classes may be valuable. I think also bioethics training is important. But to gain a more complete perspective, I think students could work as interns in labs in core facilities and they could participate in capstone projects that are designed to teach them the complete end-to end-process.
In coming up with these guidelines, are you focusing on curriculum for a specific academic level; for instance, bachelors versus masters level students?
The competencies apply at all levels. More advanced levels require a greater depth of understanding of each particular competency. For example, teamwork is needed at all levels but leadership is really needed at the more advanced levels.
According to the paper, one of the more challenging things the committee faced was trying to come up with a general way through which institutions could share details about their bioinformatics programs. Let’s talk a bit about the way you resolved that situation.
Russell Schwartz came up with the idea of developing a controlled vocabulary for describing programs. The idea is that if people would just use the terms from that controlled vocabulary then we would be having more of a standard way of comparing and describing the program.
How are you planning to disseminate these guidelines to the community?
There are several ways through ISCB. This will be made available through the ISCB website. We also have a committee of special interest focusing on education within the ISCB so they’ll be made available to that community. Also that community of special interest on education includes GOBLET [Global Organization for Bioinformatics Learning, Education, and Training] and [they’ll] also be helping to disseminate this information. Then there’ll be a Birds of a Feather (BOF) meeting at the ISMB conference in July where these will be presented and discussed and the floor will be opened up for comments.
Also, I’m participating in curriculum development efforts in several places. There’s a group in Africa that came together about two weeks ago and they are developing Masters degree programs. I presented these guidelines and I heard from the organizer of the meeting that they actually used these guidelines as some of the foundations for the curriculum that they developed. Then, I’m involved with a US biology education group and it’s going to be having a meeting out in Nebraska next month and I’ll be presenting these there. I’ve also presented a draft of these at the meeting of the American Association of Medical Colleges; they’re considering how to bring [bioinformatics and computational biology] into the (already full) medical curricula.
This is a question that’s been discussed in other papers and settings but I’d like to get your take on it as well. Why is it so difficult to establish a bioinformatics curriculum?
I think it’s due to a diversity of perspectives [and] diversity of education systems. The education systems on different continents sometimes can be quite different. There are also a number of different disciplines that are coming together here: biologists, computer scientists, mathematicians, medical practitioners, researchers, and we all bring different perspectives and speak different languages.
Is it also complicated by the fact that there are different types of bioinformaticians? For example the bioinformatics scientist versus the bioinformatics engineer both of which have a separate set of core competencies, as you point out in the paper?
Yes, it’s definitely complicated by that. And we hope that identifying these different categories will lead to substantial improvement in the dialogue and eventually in practice.
As we wrap up, aside from spreading the word about these guidelines within the bioinformatics community, what are the next steps for the committee?
This is not a finished work. We plan to continue refining the competencies. There’s also some talk about possibly developing assessments, measurements, [and] standards. At the BOF meeting in July, we will talk about next steps. We are also talking with the GOBLET community about joining forces.