Andrew Dalke started working in the computational life sciences in
1992 by co-developing a parallel version of CHARMm. He went on to
graduate school in Klaus Schulten's
group at the University of Illinois. He was part of the original
team behind VMD and
NAMD, and was the
main developer of VMD for two years.
He entered commerical software development with Look/GeneMine at
Molecular Applications Group, and went on to be the principal
developer of DiscoveryBase, a turnkey intranet application which
integrated several bioinformatics databases and analysis
tools.
Writing several thousand lines of Perl code was enough to convince him
that Python was a better language. He co-founded Biopython with Jeff
Chang, and went on to be the Secretary of the Open Bioinformatics
Foundation. He is also a member of the Python Software Foundation and
has presented at many Python conferences.
After MAG he worked at Bioreason, which used machine learning to
analyze high-throughput chemical screens. It's there where he started
in cheminformatics, and where he developed PyDaylight, a high-level
interface to the Daylight toolkit.
In 2000 he started his own consulting company in Santa Fe, NM,
primarily developing custom in-house software for early-stage
pharmaceutical R&D.; He and the company moved to Gothenburg, Sweden in
2007. One of the few publicly discussed projects is C-Lab, an internal
descriptor and model prediction web application for AstraZeneca.
Andrew developed the PyDrone back-end component, which integrates
several dozen local and third-pary components, and manages the
dependencies between them.
Between 2004 and 2007, Andrew made several visits to South Africa to
teach beginning and intermediate courses on Python for
bioinformaticians for the National Bioinformatics Network, and to give
more advanced training and advice at the University of Pretoria and
Stellenbosch University.
In 2008, Andrew move to Sweden, and started Andrew Dalke Scientific,
AB.
Most of his work since then has been in cheminformatics. Some of the
projects of note are the chemfp
high-performance similarity search package, the multi-structure
maximum common substructure algorithm now in RDKit, and the mmpdb matched molecular
pair analysis tool.