This author interview is by Dr Mark Miller, of University of California’s San Diego Supercomputer Center.  Dr Miller's full paper, A RESTful API for access to phylogenetic tools via the CIPRES Science Gateway, is available for download in Evolutionary Bioinformatics.

First please summarise for readers the content of your article
Getting access to large computing resources is key for a lot of modern phylogenetics research. Our goal is to make this access ever easier for scientists. Our original CIPRES service allows researchers to run parallel codes for sequence alignment and tree inference on a set of large compute clusters in the US. Access was provided through a web browser. Sometimes a web browser is perfect for submitting jobs, but sometimes it is tedious and interruptive. The article describes a new CIPRES web service that allows scientists to submit the same sorts of jobs outside of a browser interface.  The importance of this is two-fold.  First, anyone who knows a scripting language can now run jobs from their desktop /laptop computer. If one wanted to submit 50 jobs at once, it is now possible without clicking 50 create job buttons, and 50 submit buttons. Second, there are a lot of highly sophisticated desktop applications out there, but they can only use whatever compute power is available where they are installed. Sequence alignment and tree inference are very compute-intensive now, and take forever to complete in desktop environments. With our new web services, desktop applications run locally, but when a sequence alignment or tree inference analysis is needed, they are sent off to run on supercomputers at SDSC. The user never needs to be aware that the jobs are running elsewhere, the results come much faster, and they arrive in the work environment where they can be easily analysed.

How did you come to be involved in your area of study?
I have a PhD in Biochemistry. One day in 2000, I drove up to a Wendy's and saw a young woman taking orders on a tablet computer. I thought how much better life would be for Biologists if they had tools like that to enter their data. And that began my interest in building IT tools that help biologists get their work done.  

What was previously known about the topic of your article?
The technologies for web services have been with us for more than a decade, and their implementation in service of Biologists dates back nearly as far.  The growth in popularity of web services has recently skyrocketed, and it directly parallels the explosion of interest in mobile apps, which have web services at their core.  The basic premise of these apps is using web services to bring lots of remote compute power to a tiny device for display. Our goal is exactly the same. In our case, the work represents the first publicly accessible implementation of those technologies for the phylogenetics community.

How has your work in this area advanced understanding of the topic?
Our work opens the door to new ways of using technology in service of Phylogenetic research. What we know already is that we have increased the power of users to compute trees and sequence alignments in several web sites, and soon we expect the same will be true for several desktop applications.  We look forward to seeing how wide the adoption is, and in how many new ways these services will be adapted in new software projects.

What do you regard as being the most important aspect of the results reported in the article?
The most important thing is opening up access to high performance cluster computing to a wide audience of development-savvy computational biologists. They are now free to incorporate and adapt these services in new and creative ways to enable a new generation of phylogenetics software. It is possible now to create tools for the desktop that reach out for supercomputing power when it is required, while still providing the convenience, flexibility, and dynamic interface capabilities of a desktop app.

share on