[ ISCB ]      [ News ]      [ Forum ]      [ Wiki ]
:: MAIN MENU ::
:: Home
:: Structure
:: Newsletters
:: Forum
:: Wiki
:: History
:: Conferences
:: Contributors
:: Feedback
:: COMMITTEES ::
:: Events
:: Business Relations
:: Marketing & Operations
:: Education & Careers
:: RSGs

Back

Machine Learning

Machine Learning
Modeling substrate specificity of human dipeptidyl-peptidase III using Random Forests
Fran SupekMarija Abramić, Tomislav Smuc
Rudjer Boskovic Institute, Croatia

Substrate specificity of the human dipeptidyl-peptidase III has been modeled using Random Forests on representations of amino acids by three general physicochemical properties. Site P1 has a strong impact on peptide binding affinity, but does not affect cleavage. A hydrophobic amino acid at site P1’ is favorable for both processes.

Download PDF of full abstract.


Machine Learning
Combining dissimilarity based classifiers for cancer prediction
using gene expression profiles

Ángela Blanco, Manuel Martín-Merino, Javier De Las Rivas
Universidad Pontificia de Salamanca, Spain

Machine learning techniques allow to identify cancerous tissues using gene expression profiles. However, the techniques proposed in the literature fail to identify cancerous tissues (false negative errors) which is a serious drawback. To overcome this problem we present a new classification scheme that combines different dissimilarities to reduce particularly false negative errors.

Download PDF of full abstract.


Machine Learning
Using Gene Regulatory Networks to Study Gene Interactions in Human Liver Cancer
Ibrahim Emam, Rasha Mokhtar, Ashraf Abdelbar
The American University in Cairo, Egypt

A Bayesian network model was used to build two gene networks for a set of proliferation genes to compare gene interactions in normal liver tissue versus Human Hepatocellular Carcinoma(HCC) caused by HCV. Results provided valid biological hypotheses about tumor development in HCC, which might lead to new drug targets.

Download PDF of full abstract.


Machine Learning
Learning Yeast Functional Upstream Open Reading Frames from
Selpi, Christopher Bryant, Graham Kemp, Marija Cvijovic, Per Sunnerhagen, Olle Nerman, Erik Kristiansson, Janeli Sarv, Alexandra Jauhiainen
The Robert Gordon University, United Kingdom

While identifying functional upstream open reading frames (uORFs) is important in understanding uORF's roles in gene regulation, laboratory experiments are expensive. Our logic-based approach to predicting functional uORFs in Saccharomyces cerevisiae uses knowledge derived from sequence data, expression data, and GO annotations. The method gives 81% sensitivity, and simple-interpretable hypotheses.

Download PDF of full abstract.