Biomedical Engineering Reference
In-Depth Information
and an Opera High Content Imaging System. All our high-throughput
projects occur continuously, at time of writing we are working on 16
genomics projects concurrently, with 20 people on these projects. Projects
can run over 1-3 years (or more) being the main focus of a post-doc's
project perhaps because a new model organism needs to have a
comprehensive set of genomic resources, or it can be a small component
because we need to do a new RNAseq gene expression analysis with an
existing model organism. Typically, our workfl ows would involve going
from raw sequence reads to generating a rough draft assembly, which we
would annotate
de novo
or with RNAseq methods or using NGS sequence
to identify SNPs in both model or non-model species, which would be
verifi ed in the laboratory and developed into markers for disease
resistance. We use our new models in genomics analyses including
comparative genomics, the genes we predict will be classifi ed and certain
important domains and signatures identifi ed to create candidates that
may confer disease or resistance to disease for further lab testing.
The hardware we use is suffi cient to cover the needs of a group of
our size, given our wide interests. Our computer systems comprise
21 TB 'live' storage on a mirrored NetApp [5] device (aside from a
substantial archive for older data) with 22 compute nodes with 2-16
processor cores each and from 8 GB up to 128 GB onboard RAM, these
all run Debian 6.0. Our bioinformatics service has maintained its small
size throughout this period of expansion and our raison d'etre is to
provide the rest of TSL with the knowledge, expertise and facility to
carry out the bold research that is their remit. Our philosophy on how we
carry out this provision has shifted over this period, from being a
primarily reactionary on-demand service to being a scalable, department-
wide initiative to create a culture of informatics knowledge, that is to
bridge the somewhat false gap between the bench scientist and the
bioinformatician.
We consider that a successful bioinformatics team helps the department
best by ensuring that whatever the biological question or source
of the data we are able to continually understand, manage and
analyse the data. We take responsibility for providing the bench
scientist with the skills and tools to carry out analyses much more
independently than in a reactionary service model. We have found
that we can execute the bioinformatics work for at least as many
projects as we have scientists (not just as many as can be handled by
two core informaticians). In this context a large part of the job
comes in facilitating that exercise. Partly this is through training - much
in the manner that a post doc would look after a new graduate
Search WWH ::
Custom Search