Inside Genomics: Q&A : Brian Haas of the Broad Institute

Abstract

NARGES BANI ASADI, Ph.D.

One of the elements lacking in the personalized medicine discussion today is the perspective of leading clinicians, informaticians and academics working in the field. To remedy the gap, I've asked a series of industry leaders to offer up their views.

I recently spoke with Brian Haas at the Broad Institute, who gave a talk on more efficiently assembling RNA-Seq data via a massively parallel computing architecture.

With degrees in both molecular biology and computer science, Brian sits squarely at the intersection between biology and high performance computing. He is the lead developer of the Trinity RNA-Seq de novo assembly software and has over a decade of experience in genomics and bioinformatics. He worked at The Institute of Genomic Research/J. Craig Venter Institute for eight years before joining the Broad Institute, where he has co-authored numerous scientific publications in leading journals and developed several popular Open Source bioinformatics tools that are widely used by the genomics community for genome and transcriptome research.

Here's our conversation:

Dr. Asadi: What are the biggest problems that you think fast and accurate genomics analysis can solve in the future? Is it cancer? Or something equally ground-breaking?

Haas: Cancer is certainly one of the prime targets for clinical applications, as knowing the potential drivers should in many cases guide more tailored therapies. In addition, accurate and fast sequencing and analysis is destined to enable ground-breaking discoveries in nearly all areas of biology. Every living organism on our planet is a potential target. Between genome, transcriptome, and epigenome analysis, and being able to assay sequences at cellular or even sub-cellular resolution, the amount of data to be generated and analyzed in the future and the potential for revolutionary discoveries is mind-blowing.

Dr. Asadi: In your opinion, what is the biggest factor that will drive medical breakthroughs via bioinformatics? Better data access and shareability? Faster NGS analysis? The identification of more variants and their biological implications?

Haas: Better data access and data sharing is going to be a major driving factor, and there are very serious barriers to this for which many groups are exploring potential solutions.

That said, within the next ten or twenty years, I think most everyone will have their genome sequenced. It will likely be a routine procedure for newborns, where the genome sequence becomes just one part of your electronic medical records. This, coupled with rich metadata stored over a lifetime, should empower many great discoveries and ultimately improve patient care. At the same time, the potential for discrimination and breaching of privacy is a major cause of concern, and one of the major focal points for bioinformaticians today.

Brian Haas

Dr. Asadi: How have you seen the genomics data analytical and storage requirements change over the last few years? What general solutions to these challenges are you all hoping to see?

Haas: The exponential increases in the yields of modern sequencing instruments has translated into an explosion in storage requirements and an ever-growing need for high performance computing systems to be leveraged for processing and analyzing these data.

Cloud computing (such as through Amazon) seems to be a popular option, particularly to those that do not want or need to invest in their own hardware. Since both cost and data privacy are major concerns, some research groups decided to build their own private clouds with significant cost savings, leveraging commodity hardware and open source software.

It's been great to see how many hardware and software vendors are teaming up with academic research groups to help tackle challenges in genomics, and best leverage the latest advancements in computer processors and hardware configurations. That's a trend that I expect will continue.

Dr. Asadi: If you could do it all over again, what would you have done differently in your career and/or research to date, so as to have a bigger impact today?

Haas: Given how I spend most of my time writing code and doing bioinformatics, it certainly would have helped to have taken more computer science and mathematics classes back in my formative years. I'd like to think that my years of pipetting samples, pouring gels, etc., set me on the path to where I am today, and just a handful of mathematics classes shy of where I ought to be.