[Solomonov Seminar] [E8] 209. Solomonov seminar

[Marko Grobelnik]

V petek 4. septembra bo ob 10:00h v Oranzni predavalnici IJS
(drugo nadstropje glavne zgradbe), 210. Solomonov seminar.
Posnetki preteklih seminarjev so na http://videolectures.net/solomon/

Predaval bo nas partner na vec projektih Juho Rousu iz Finske 
na temo strukturnega ucenja na podrocju bioinformatike.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Structured Output Prediction of Enzyme Function via Reaction Kernels

Juho Rousu
Department of Computer Science
University of Helsinki, Finland

Enzyme function prediction is an important problem in post-genomic 
bioinformatics. There are two general methods for solving the problem: 
transfer of annotation from a similar, already annotated protein, and
machine learning approaches that treat the problem as classification 
against a fixed taxonomy, such as Gene Ontology or the EC hierarchy. 
These methods are suitable in cases where the function has been 
previously characterized and included in the taxonomy. However, given a 
new function that is not previously described, existing approaches 
arguably do not offer adequate support for the human expert.

In this presentation, we I will present a structured output learning 
approach, where the enzyme function, an enzymatic reaction, is described 
in fine-grained fashion with so called reaction kernels which allow 
interpolation and extrapolation in the output (reaction) space. A 
structured output model is learned to predict enzymatic reactions from 
sequence motifs. We bring forward several choices for constructing 
reaction kernels and experiment with them in the remote homology case 
where the functions in the test set have not been seen in the training 
phase. Our experiments demonstrate the viability of our approach.