David KulpResearch Associate
Steve McCarroll's Neurobiology Lab
Department of Genetics
Harvard Medical School
EducationPh.D.(2003), Computer Engineering (Bioinformatics), UC Santa Cruz
Dip.Sci.(1992), Computer Science, University of Canterbury, New Zealand (Fulbright Scholar)
B.Sci.(1991), Computer Science, College of William and Mary
I am a research associate at Harvard Med specializing in bioinformatics. Currently my focus is on application development for visualization and analysis of data derived from single cell sequencing technology. My most recent public work was the release of DropViz, a tool to explore an atlas of the mouse brain (Cell; 174(4)).
Previously I worked as an independent consultant and software developer and before that I was on the faculty of the Computer Science department at the University of Massachusetts in Amherst.
I left university during my tenure track when my first wife died and I became a single dad. After my departure from UMass I consulted off and on in biotech — I worked on iOS development building a prototype mobile genomics app for the UCSC Genomics Institute and with Ion Torrent refining their base-calling algorithm. I briefly jumped into a completely new area—online advertising—where I worked for two years on "big data" in the cloud crunching numbers to develop personalized models used in Real-Time Bidding.
I also occassionally collaborate with Andrea Foulkes at Mt Holyoke on methods for whole genome association studies.
My work has historically been in bioinformatics. My interests were concerned with how DNA variation and mutations are associated with basic molecular changes in cell biology and disease. To this end, I engaged in research concerned with technology improvements for detecting single feature polymorphisms using microarrays; genetic linkage models for analyzing variation data and molecular-level (gene expression) quantitative traits to derive pairwise interactions among genes and local regulatory networks; and discovery programs applying variation detection assays to microbial and viral genomes.
I took a long strange hiatus from university in 1997 to make a go of a bioinformatics start-up with some buddies of mine in Berkeley. We called it Neomorphic. We mostly did custom software development (LIMS for high-throughput sequencing, visualization toolkits, genome browsers) and I worked on research collaborations and sequence analysis algorithm development.
At Neomorphic, I was involved with two amazing international collaborations to analyze the genomes of Drosophila melanogaster and Homo sapiens. In both cases, I applied the Genie software to the entire genome, and, in the case of the human genome, performed additional whole-genome analysis about the draft assembly and the general characteristics of genes. I finally got around to putting the details of the current system into a manuscript.
Despite many struggles, Neomorphic was profitable and successful. In mid-2000, we merged with Affymetrix in a great marriage of bioinformatics strength and microarray leadership. I managed the bioinformatics group with wide-ranging responsibilities including research and production processes in chip design, probe modeling, genome annotation, functional prediction, genotyping and expression signal analysis, comparative genomics, splice variant detection, etc. It is an amazing company doing incredible science. In 2002 my group published on the annotation frameworks used at Affymetrix for functionally characterizing genes represented on commercial GeneChip microarrays. In 2003 the initial analysis of the mouse genome was published and my group published on alternative splice detection with microarrays, probe selection, and genotyping algorithms.
I worked for about a year for Live Picture, a long gone software company that developed software around their multi-resolution JPEG format, which became FlashPix. About the time I got there, they were making an early internet play and trying to jump into the consumer digital photography market (about 5 years too soon). I wrote the server, helped with the client plug-in, and penned the protocol for navigating through high-resolution images.
Before I moved to Santa Cruz, I worked at the Genome Database at Johns Hopkins in Baltimore from '93 to '94, just as the World Wide Web and Mosaic were hitting the internet. I wrote the front-end to "GDB v5" as a web interface that allowed systematic browsing and data retrieval throughout a complex bioinformatics database. I believe it was the first of its kind in those early days of the web. I wrote it using Perl and CGI.
Before moving to Baltimore in the early '90s, I was a vagabond traveler on a slow trek back to the states from New Zealand. I had spent over a year in graduate school at the University of Christchurch, New Zealand as a fulbright student. I had no clue what I was doing and thought I wanted to pursue environmental simulations, but as luck would have it the prof I was going to study with went on sabbatical and I ended up studying implementation efficiencies for lossless compression with Tim Bell - a great guy and co-author of two excellent primary references in the field. In fact, thanks to Tim I had the good fortune to end up at UC Santa Cruz where the famous Huffman and arithmetic coding both have their roots.
I completed my undergraduate studies in Computer Science at the College of William and Mary in Virginia. I minored in philosophy. From an academic perspective, college was completely unremarkable except for a freshman writing seminar that I still remember well. My best memories were DJing on WCWM, jamming with Steve Swallow and Carla Bley, playing drums in a funk/rock cover band we called Budnuda, and making some life-long friends and meeting my first wife, Laura.