Learning to recognize brain specific proteins based on low-level features from on-line prediction servers

During the last decade, the area of bioinformatics has produced an overwhelming amount of data, with the recently published draft of the human genome being the most prominent example. This has enabled researchers to use data driven, rather than hypothesis driven, methods to address a wide variety of specific problems related to the analysis of biological sequences (e.g., protein, DNA and RNA sequences). Today a number of low-level properties of biological sequences, like the presence or absence of signal peptides, can be obtained from publicly available on-line prediction servers. Such a server typically implements a classifier which is trained to determine a single property of a sequence on the basis of various kinds of biochemical laboratory results. In this paper we investigate how the low-level data from these distributed on-line sources can be combined to construct a classifier that recognizes a high-level property, namely the brain specificity, of a protein. This is a task for w...

• Type: Conference
• Year: 2001
• Where: KDD
• Authors: Henrik Boström, Joakim Cöster, Lars Asker, Mikael Huss
• Keywords: Brain Specificity, Data Mining, KDD 2001, Keywords Brain Specificity, On-line Prediction Servers