PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Extracting yeast stress genes by dependencies between stress treatments
Arto Klami, Janne Nikkilä, Christophe Roos and Samuel Kaski
In: 4th European Conference on Computational Biology, 28 Sep - 01 Oct 2005, Madrid, Spain.


It is much harder to define stress response than it is to name stressful treatments. This inspires a data-driven definition of stress as a response shared by multiple stressful treatments. More specifically, we assume that statistical dependencies between stress data sets are due to stress. We search for environmental stress response genes in yeast with a new non-parametric method that detects dependencies in data sets with co-occurring samples. We introduce a method for detecting and discarding the samples that are explained by the individual sets. The remaining samples are then ordered according to how badly the individual sets explain them. The ordering is based on non-parametric estimates of specific log-ratios, and in particular high log-ratios are likely to be signs of dependencies. The potential yeast stress genes were here sought based on the expression data from several (17) stress treatments (from [1] and [2]): heat (3), acid, alkali, peroxide, NaCl, sorbitol (2), H2O2, menadione, dtt (2), diamide, hypo-osmotic, aminoacid starvation, and nitrogen depletion. A short time series had been measured from each, and in total we had 108 dimensions for 6013 yeast genes. Ordering the genes revealed the existing yeast stress genes with good accuracy: 804 out of 868 environmental stress response genes found in [2] were found among the group having high log-ratios. The method also identified a set of new genes that can have a role in stress response. About half of the 97 new stress genes studied more closely have already been attributed other primary functions than stress response, but our analysis suggested that these genes could also be involved in environmental stress response. Furthermore, 25 genes were of unknown function, and our classification thus gave a new hint of in what type of processes they could be involved. References: [1] Causton et al (2001) Remodeling of yeast genome expression in response to environmental changes, Mol. Biol. Cell, 12, 323-337. [2] Gasch et al (2000) Genomic expression programs in the response of yeast cells to environmental changes, Mol. Biol. Cell, 11, 4241-4257.

PDF - Requires Adobe Acrobat Reader or other PDF viewer.
EPrint Type:Conference or Workshop Item (Poster)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Multimodal Integration
Theory & Algorithms
ID Code:1493
Deposited By:Arto Klami
Deposited On:28 November 2005