During the sixth annual Summit on Translational Bioinformatics held this week in San Francisco, participants were treated to talks and posters covering applications such as pharmacogenomics informatics, data mining to identify drug-drug interactions and phenotypes, and efforts to integrate data to improve disease classification (see related story this issue).
Also up for discussion were ethical, legal, social, and privacy issues that need to be taken into account as genomics data continues to proliferate and move from research into clinical settings such as gene patenting and questions of ownership.
After the meeting, BioInform sat down with Jessica Tenenbaum, who chaired the scientific committee for TBI 2013, to talk about this year's conference and the current state of the translational bioinformatics field. Tenenbaum is the associate director of bioinformatics at Duke Translational Medical Institute's Biomedical Informatics core. What follows is an edited version of the conversation.
What were some highlights of this year’s meeting for you?
I was really pleased with how the posters in the “Silver Medallions” session came off — that was new. In the past, we’ve tried to incorporate posters into the program with poster bullets and it didn’t work well in this environment. I was thrilled with how the Silver Medallion panel contextualized the posters and with the audience response … there was a great discussion.
The Silver Medallion session was something new. It was meant to be senior people in the field who have a big-picture view choosing selected posters and framing them in the context of the field to explain what the bigger picture is and how that work fits into it. That worked out really well.
The ethical, legal, and social issues panel where we were discussing gene patenting and return of results and ownership of gene information was a great discussion. And also the podium presentations … tend to be outstanding. I only got to attend some of the scientific sessions this year but I was so pleased with the quality of the work that I saw and the relevance to other people’s work.
It sounds like you’ve been coming to the TBI meeting for some time. How has the conference, and, more generally, the translational bioinformatics field progressed since those early days?
[TBI’s] been going on six years and I’ve been here since the beginning. They used to have pre-assigned tracks and … typically one of the tracks was a catch-all of methods in translational bioinformatics. Eighty percent of people would submit to that and we’d have to try and separate it out. This year, we tried something different. We just had people assign keywords and then did a post-hoc clustering of what the different tracks were. I would say that the kinds of themes you see have changed. Now we are seeing a lot of pharmacogenomics and cancer informatics [which are] sessions that you wouldn’t have seen in the first one. Also text mining, natural language processing methods, and mining electronic medical records.
I think at the beginning it was still more the algorithms, the ways we could do this, and over the last two years you are starting to see a lot more papers about something that happened in the clinic, something that came to the patient level rather than remaining a proof of concept that shows we can discover a gene.
What challenges would you say still need to be addressed?
There is a lot of focus on the genomic data. Metabolomics and proteomics are further downstream, arguably more relevant but still a little bit of the Wild West. They are both in some ways harder to do; the standards are less well-developed so that’s definitely a challenge. I think we’ll start to see more of those.
Next-generation sequencing is moving along very fast … it’s the usual problem where there are a lot of different groups doing it. There are standards at some level [set by] the Broad Institute and some of the other large sequencing centers but hopefully we’ll see some of the groups coming together to make sure that we are moving forward in a united way.
Following up on that, I noticed that there were a lot of presentations around other types of data such as gene expression data and electronic medical records but not too many that used next-generation sequencing data. Why do you think that is?
I think we’ll see more of that. There are more and more publically available datasets out there but I think the results of those tend to be some of the bigger papers that are presented in journals and less so at conferences. Or if they are at conferences, [they are] at the American Society of Human Genetics. Next year’s chair, Joshua Denny [an assistant professor of medicine and biomedical informatics] from Vanderbilt [University], is involved with ASHG — hopefully we’ll have more crossover with that group next year. I firmly believe there are a lot of people out there who are translational bioinformaticians who are doing TBI and don’t know it. Someone in one of the sessions made the point that people at ASHG are tackling all these problems that this group has already … if not solved … gone a long way to addressing, and we need to bring those communities together.
What would you hope to see from the field of translational bioinformatics say five years from now?
I hope that we would have made some progress in what I see as the biggest challenge; this idea that we’re getting closer to the $1,000 genome but the $100,000 analysis. Hopefully, we as a field will help make that a little more manageable. That’s the main thing.