The nature of statistical learning theory
The nature of statistical learning theory
Making large-scale support vector machine learning practical
Advances in kernel methods
Gene functional classification from heterogeneous data
RECOMB '01 Proceedings of the fifth annual international conference on Computational biology
Transductive Inference for Text Classification using Support Vector Machines
ICML '99 Proceedings of the Sixteenth International Conference on Machine Learning
Using the Fisher Kernel Method to Detect Remote Protein Homologies
Proceedings of the Seventh International Conference on Intelligent Systems for Molecular Biology
Iterative Weighting of Phylogenetic Profiles Increases Classification Accuracy
ICMLA '05 Proceedings of the Fourth International Conference on Machine Learning and Applications
| Hi-index | 0.00 |
Phylogenetic profiles of proteins - strings of ones and zeros encoding respectively the presence and absence of proteins in a group of genomes - have recently been used to identify homologous proteins and/or proteins that are functionally linked, such as participating in a metabolic pathway. We proposed a novel learning method for protein classification based on phylogenetic profiles, which takes into account both the phylogenetic tree structure and the likelihood of proteins presence in genomes. The method consists of a mechanism to extend the profiles with extra bits encoding the phylogenetic tree, whose interior nodes, representing hypothetical ancestral genomes, are scored in a way to reflect their chances of developing divergence in the descendants. The scoring scheme also incorporates the likelihood of proteins presence in genomes as weighting factors, which are collected from the training data initially and integrated as part of kernel of a support vector machine. In a transductive learning scheme, when the SVM is used for classifying test data, the weighting factors are updated iteratively using the predicted results. We tested our method on the proteome of Saccharomyces cerevisiae and used the MIPS classification as a benchmark. The results showed that the classification accuracy was greatly increased.