chris
Email: email
CV: pdf | odt | doc
{ newsflash }
As of Friday 23rd February 2007, I will no longer be working in Liisa Holm's group. I begin a new post in bioinformatics support at the Babraham Institute, Cambridge, England, on Tuesday 6th March 2007. If you want to contact me there, my email address will be:
chris.wiltonREMOVE_THIS_BIT@bbsrc.ac.uk
Amongst other achievements, I am proud to have redesigned these very web pages to give the Holm Group website a more professional appearance. All our pages should be XHTML-1.0 valid, and CSS-2 valid, and should display the same in all browsers. The menubar is CSS-only, no javascript involved at all.
Wanna know what the old pages looked like? Take a look here...
{ current research interests }
I am currently involved in parallelisation of the GTG build/update procedure, for running on the DEISA Grid. We have found that our production server simply cannot cope with the power and memory resources needed to calculate the graph, due to the ever-increasing number of known protein sequences - it would take us more than six months to calculate the new version! Therefore, we made an application to the DEISA initiative, which was subsequently granted (thanks to Shop, Andreas, Liisa and the CSC team). In order to make use of the DEISA grid, we now need to parallelise the procedures for use on upwards of 64 processors, before we can run the first update.
The Global Trace Graph (GTG) has been developed from an original idea by Liisa Holm and Andreas Heger, for a novel transitive alignment system based on consistency scores - in very basic terms, we can score an alignment between any two residues in two proteins by looking at the size of the overlapping neighbourhoods of the two proteins compared with their combined neighbourhood size. This puts a value on our confidence in any aligned residue pair. We can then trace paths through the entire graph to transitively connect any pair of residues and assign that connection a score. The GTG is a pre-computed, pruned, all-against-all alignment graph of a representative set (40% identical) of all known proteins.
The new GTG Server is ready. Use it for fold recognition, for finding distant homologues, for retrieving alignments as a base for homology modelling. The GTG-DEEP algorithm was used in our automated server for the recent CASP-2006 protein structure prediction competition, with encouraging results in fold recognition. Publication on GTG pending.
The new Pfam-Squared Server is ready. It uses the GTG to perform an all-against-all alignment of Pfam families - you can use it to investigate relationships, to discover common motifs and to obtain multiple alignments between Pfam families and clans.
I am also involved in a project to design and implement a novel hierarchical protein functional classification scheme, using evolutionary data to correctly cluster proteins into homologous super-family and functional family groups, and to assign correct functional GO annotation to them. This should overcome the major problems with existing domain classification schemes - incorrect assignment of functional annotation to a protein, due to fused genes and promiscuous domains along with the problem of a lack of consistency in gene product annotation; and potential missing of functional residues and misclassification of proteins due to strict domain-cutting rules. See PMid: 16390875.
For the third year running, I helped to lead a three-week course on 'introductory protein analysis' in November 2006, involving the preparation of lectures, demonstrations and practical exercise sessions in a number of areas including sequence analysis, multiple alignments (and manually adjustment of), homology modelling, and protein domain family classification. Whilst I enjoy lecturing, I never knew preparing all the material could be so stressful! Now I know why some people recycle their lecture material year after year...
I am using C++, Java, Python and Perl languages. For web development, I work mostly in the Zope Application environment. I hand-code all my own (valid) XHTML and CSS - no wysiwig editor comes close. Some day I'll get round to developing a decent html editor!
I also have the following responsibilities:
  • Update and maintenance of various group software, database updates, MySQL administration
  • Development and maintenance of dynamic web applications, plus the group's static html pages
  • Administration and maintenance of the group's application servers (Zope, Apache)
{ selected articles }
Masters Thesis (2003):
{ publications pending }
Information soon on Global Trace Graph and further Module Space publications.
{ previous positions }
2000 - 2002: Academic Department of Clinical Oncology, University of Nottingham, England.
Working on a vaccine against Osteosarcoma, Colorectal and other cancers that over-express the surface antigen CD55 (Decay Accelerating Factor), and testing in clinical trials. Investigating the use of Dendritic Cells for efficient antigen presentation in vitro. For more information on the continuing research, please visit Lindy Durrant's group website.
{ contact me }
Chris Wilton
Structural Genomics Group
Institute of Biotechnology
PO Box 56
00014 University of Helsinki
Finland
tel: +358 919159114
my email