I am currently a computational scientist at LBNL and the Virtual Institute for Microbial Stress and Survival (VIMSS), in Adam Arkin's group. My main interests are molecular sequence and structure evolution, with the applied purpose of gleaming details of known molecular features as well as identifying novel ones. I express these interests with computational and statistical approaches. During my Ph.D. work at UCSF in Fred Cohen's group, I built a multivalent graphical browser called Jevtrace, implementing the Evolutionary Trace method and variations (which have since been dubbed phylogenomics) using the dynamic combination of phylogeny, multiple sequence alignment, and molecular structure. Structure viewing capability came from integration with WebMol, a fast and robust molecular structure viewing and analysis tool (though it's no PyMol). Jevtrace and underlying methods were published in Genome Biology, back when it was the first fully online and open access journal. There is nothing quite like being able to present publications and unlimited color graphics online to everyone. The software proved helpful for an antimalarial drug design project and is now used by a few hundred researchers on six continents. Jevtrace is under active development and a major overhaul has just been released allowing work with large datasets, tree output from most phylogeny software, and other goodies. The software is distributed as a single executable JAR file, which in my mind revolutionizes academic software distribution: no installation is required and program execution happens with a mere mouse double-click.

• Screenshot of Jevtrace in action:
  

• Screenshot of Web Mol in action (with Jevtrace):
  Web Mol in action (with Jevtrace)" width="700" height="550" />

As a postdoc in Steven Brennner's group at UC Berkeley I worked on a number of projects. The highlight was the sequence analysis survey we performed of the protein sequences, domains, and families from the Sorcerer II Global Ocean Sampling expedition ( [J. Craig Venter Institute link). The publication appeared in PLOS Biology.

Another exciting project was the statistical modeling of human splicing signals, in collaboration with the groups of Alan Frankel and Don Rio. A computational biologist cannot survive without experimental validation and I am lucky to have such outstanding partners on the wet lab side of things. Two computational tools that have been instrumental parts of the analysis are the profile search tool motifBS and graphical sequence logo generator WebLogo.

I have also been analyzing the relationship between multiple sequence alignment variation and phylogeny reconstruction - especially the phenomenon of guide tree imprinting observed in progressive alignment by James A. Lake in 1991. This project has been in collaboration with two fabulous UC Berkeley undergraduates Jian Wang and Kang Wang. A manuscript is in final stages of preparation and a link will be posted soon.

Of interest to even the general public is a Java program called JColorGrid I wrote to generate color-grid graphics from tab-delimited text data and Excel spreadsheets. Two existing Java APIs provided some very nice functionality: parsing and generation of MS Excel files with Andy Khan's JExcel API, and generation of EPS graphics with Paul Mutton's EPS Graphics2D API. My faith in Java has faltered at times, but with the widespread availability of increasingly sophisticated APIs, generation of new software can be quite pleasant if you are lucky and can find what you need. JColorGrid has been in collaboration with the UCSF experimentalists Barney May and Jennifer Weismann.

• jcolorgrid.jpg: