Presentation 2006-06-15
Predicting type of protein-protein interaction as a multiple-instance learning problem
Hiroshi YAMAKAWA, Yoshio NAKAO, Koji MARUHASHI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We propose a method for predicting types of protein-protein interactions using a multiple-instance learning (MIL) model. By assuming cause of each interaction between proteins containing two or more subunit is results from local subunit pairs, we formulates this problem as MIL. In this problem, influences from instances in negative bag to target concepts are too strong by using well-known Maron's method. We propose a new MIL method based on decision by majority and apply to the KEGG interaction data. In an experiment using the KEGG pathways and the Gene Ontology, the method successfully predicted an interaction type (phosphorylation) at the accuracy rate of 86.1%. We find that cause of false positive is caused by positive bias peculiar to MIL method by analyzing prediction results.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) protein-protein interaction / multiple-instance learning / diverse density / phosphorylation / subunit
Paper # NC2006-15
Date of Issue

Conference Information
Committee NC
Conference Date 2006/6/8(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Predicting type of protein-protein interaction as a multiple-instance learning problem
Sub Title (in English)
Keyword(1) protein-protein interaction
Keyword(2) multiple-instance learning
Keyword(3) diverse density
Keyword(4) phosphorylation
Keyword(5) subunit
1st Author's Name Hiroshi YAMAKAWA
1st Author's Affiliation FUJITSU LABORATORIES LTD.()
2nd Author's Name Yoshio NAKAO
2nd Author's Affiliation FUJITSU LABORATORIES LTD.
3rd Author's Name Koji MARUHASHI
3rd Author's Affiliation FUJITSU LABORATORIES LTD.
Date 2006-06-15
Paper # NC2006-15
Volume (vol) vol.106
Number (no) 101
Page pp.pp.-
#Pages 6
Date of Issue