Tutorials are free of charge, but registration is required, either through the registration form, or by sending email to the conference contact
Tutorial lectures on selected topics in Bioinformatics will be held at the Computer Science Department of the University of Pisa on June 22nd in the afternoon.
Lecturer 1: Prof. Giorgio Valentini (Università di Milano).
Title: Machine learning methods for gene function prediction
Lecturer 2 : Dr. Andrea Bracciali (University of Stirling)
Title: Formal Models in Systems Biology
In this tutorial we introduce the gene function prediction problem as a complex classification problem characterized by several challenging issues, such as the large number of functional classes, the multiple annotations for each gene/gene product, the hierarchical relationships between functional classes, the different annotation evidence, the availability of multiple sources of complex and noisy data, the lack of a univocal definition of negative examples.
The first attempts to computationally predict the function of genes or gene products were based on algorithms able to infer similarities between sequences. More recently several machine-learning based methods, able to exploit multiple sources of data from high-throughput biotechnologies have been proposed and applied to this challenging problem.
We briefly overview the main machine learning-based research lines on this topic, ranging from network-based label propagation methods, kernel methods for structured output spaces and hierarchical multi-label ensemble methods.
In particular we focus on hierarchical ensemble methods outlining their effectiveness and limitations, in the context of the Gene Ontology and FunCat taxonomies.
- I. Friedberg, Automated protein function prediction-the genomic
challenge, Brief. Bioinformatics, vol. 7, pp. 225-242, 2006.
- Z. Barutcuoglu, R. Schapire, and O. Troyanskaya, Hierarchical
multi-label prediction of gene function, Bioinformatics, vol. 22,
no. 7, 2006.
- S. Mostafavi, et al. GeneMANIA: a real-time multiple association network integration algorithm for predicting gene function. Genome Biology, 9(S4), 2008.
- A. Sokolov and A. Ben-Hur. Hierarchical classification of Gene Ontology terms using the GOstruct method. Journal of Bioinformatics and Computational Biology, 8(2), 2010.
- G. Valentini, True Path Rule hierarchical ensembles for genome-wide
gene function prediction, IEEE ACM Transactions on Computational
Biology and Bioinformatics, vol.8 n.3 May/June 2011.
Abstract: Papers like "Protein molecules as computational elements in living
cells" [Bray, Nature 1995] and "Cells as computation" [Regev and
Shapiro, Nature 2002] have put forward the idea that many aspects of
living systems have a computational nature. Specifically, the
complex network of interaction and information exchange that occurs
within the biochemistry at the inter and intra cellular level, can
be assimilated to the functioning of a distributed, interactive
computational system. In the words of Bray, proteins are
"functionally linked ... into biochemical 'circuits' that perform a
variety of simple computational tasks including amplification,
integration and information storage".
Under this perspective, it has appeared natural to employ the
techniques used to model and analyse interactive computational
systems to the realm of living organisms. Such a research trend aims
at further developments within Systems Biology, the research area
that approaches the study of the living organisms at a systemic
level (see "Systems Biology: a brief overview" [Kitano, Science> 2002]).
Computationally inspired formal models and analysis
techniques are being used to carry out "in silico" experiments,
which may often represent a cheaper, faster, more ethical, more
easily measurable, and less constrained complement to the more
traditional "in vitro/vivo" investigation.
This tutorial will briefly survey some of the formal techniques,
particularly those originated from concurrency theory, which have
been adopted, adapted and further developed for
the research in Systems Biology. Starting from a historical
perspective, the main ideas of the approach will be discussed and a
few small examples practically worked out.