Ian Holmes

Contact details

• Email: ihh berkeley edu
• Skype: ianholmes
• AIM/iChat: dartcoder@mac.com

• Office address: 374C Stanley Hall, UC Berkeley, CA 94720-3220
  • Office phone: (510) 666 - 2790
  • Department: Bioengineering
  • Office hours: 11am-noon Wednesdays, 1-2pm Fridays

• Lab address: 381 Stanley Hall, UC Berkeley, CA 94720-3220
  • Lab phone: (510) 666 - 2791

Affiliations

I am on the faculty of the Department of Bioengineering in UC Berkeley and the Physical Biosciences Division at Lawrence Berkeley Laboratory.

I'm affiliated with the following UC Berkeley/UCSF Graduate Groups:

• Computational Biology (UCB)
• Bioengineering (UCB/UCSF)
• Biophysics (UCB)

I am on the scientific advisory board of CLC Bio and the board of the Evolutionary Software Foundation.

Online locations

• Social networking sites: FriendFeed (LinkedIn, Facebook, Twitter, Last.fm ...)
• Genome Dynamics, my teaching blog
• Literature references (bibtex repository browser)

Research interests

Short summary:

• Using computers to investigate genomes, their evolution and ecology.
• Developing the statistical bioinformatics tools for this, e.g.
  • phylogenetic alignment;
  • stochastic grammars, string transducers and tree transducers;
  • multiple alignment of RNA;
  • transposon classification.
• Developing "Web 2.0" genomics infrastructure, e.g. AJAX genome wikis.
• Synthetic biology, especially RNA engineering and paleo-genetics.

Genome evolution

Computational biology needs realistic, predictive, quantitative models of the how biological sequences --- and systems --- evolve.

My lab develops & applies stochastic process models (e.g. discrete-state continuous-time Markov chains) for the study of molecular evolution.

The dynamics we have modeled include substitutions, insertions and deletions ("indels"), microsatellite dynamics, local duplications, inversions, transpositions, recombinations and rearrangements.

Systems we have considered include sequences, gene families, cis-regulatory networks and chemical signaling pathways.

Estimating missing data

Statistical inference tasks in compbio are often inter-related. For example, in order to simulate evolution, so as to mimic real sequence alignments, you'll need estimates of the underlying substitution rates from real sequence data -- which you can only measure if you've built alignments & trees.

Typically some answers can be estimated, while others need more precise measurements. Probabilistic models can provide a common framework for all the questions, reconciling parallel approaches to the same problem or suggesting new approaches.

Reconstructing the origins of life

I am especially interested in using such models to make inferences about the origins of life by working backwards from present-era DNA sequences, with the goal of reconstructing those origins in the laboratory synthetically.

I am also interested in predictive applications of the models such as genefinding, SNP analysis, and directed evolution.

Genome ecology

A related interest is "genome ecology": the (evolutionary) interactions of genomes with their neighbors.

Examples include the bioinformatics of transposon classification and the metagenomics of microbial communities.

For more info see the front page, the Holmes lab page or the paper archive.

Biographical info

• My NIH biosketch (MS Word doc)
• My publications

I grew up in Cambridge (UK) and studied physics at the Cavendish Laboratory (TCM group) and genomics at the Sanger Institute (Informatics).

I now work at UC Berkeley and live in East Oakland.

Computer games

Before I went to college, I wrote a couple of computer games:

• Pipeline, co-written with William Reeve, was published in 1989 by Superior Software for the BBC Micro. It rates a mention on the Wikipedia page for Jupiter's moons in fiction. Essentially it was a fast four-way scrolling arcade/puzzle/RPG with a level designer. There's a review here, with a download you can run on an emulator. A PC version, Pipeline Plus, is available from Superior Interactive.

• Galactic Dan, with graphics by James Davidson, was published in 1992 by Fourth Dimension for Acorn's ill-fated Archimedes. Here's a clip on Youtube. An early, cheap & cheerful first-person shooter/platformer, it apparently "pales into insignificance" next to Wolfenstein; but some people liked it. In retrospect, the tone I was aiming for was something like Duke Nukem. Dan re-appears on several compilations, lately Shootup Games Collection from APDL (review). You can find the cheat codes online.

Here are some (mostly retro) games that I've enjoyed over the years:

• Exile: physics, emotion, synthetic biology, Faustian hubris... all in 32,768 bytes
• Core Wars because programming should be Darwinian
• Shades. Old-school 1980's multiuser dungeon on Prestel
• The Sentinel and Cholo: spooky early 3D
• Sutwin, Tangle, Shade & anything else by Zarf (Interactive Fiction!)

Miscellany

• Older links from this page