Tala M Bakheet1, Andrew J Doig. 1. Faculty of Life Sciences, The University of Manchester, Michael Smith Building, Oxford Road, Manchester M13 9PT, UK.
Abstract
MOTIVATION: We analysed 148 human drug target proteins and 3573 non-drug targets to identify differences in their properties and to predict new potential drug targets. RESULTS: Drug targets are rare in organelles; they are more likely to be enzymes, particularly oxidoreductases, transferases or lyases and not ligases; they are involved in binding, signalling and communication; they are secreted; and have long lifetimes, shown by lack of PEST signals and the presence of N-glycosylation. This can be summarized into eight key properties that are desirable in a human drug target, namely: high hydrophobicity, high length, SignalP motif present, no PEST motif, more than two N-glycosylated amino acids, not more than one O-glycosylated Ser, low pI and membrane location. The sequence features were used as inputs to a support vector machine (SVM), allowing the assignment of any sequence to the drug target or non-target classes with an accuracy in the training set of 96%. We identified 668 proteins (23%) in the non-target set that have target-like properties. We suggest that drug discovery programmes would be more likely to succeed if new targets are chosen from this set or their homologues.
MOTIVATION: We analysed 148 human drug target proteins and 3573 non-drug targets to identify differences in their properties and to predict new potential drug targets. RESULTS: Drug targets are rare in organelles; they are more likely to be enzymes, particularly oxidoreductases, transferases or lyases and not ligases; they are involved in binding, signalling and communication; they are secreted; and have long lifetimes, shown by lack of PEST signals and the presence of N-glycosylation. This can be summarized into eight key properties that are desirable in a human drug target, namely: high hydrophobicity, high length, SignalP motif present, no PEST motif, more than two N-glycosylated amino acids, not more than one O-glycosylated Ser, low pI and membrane location. The sequence features were used as inputs to a support vector machine (SVM), allowing the assignment of any sequence to the drug target or non-target classes with an accuracy in the training set of 96%. We identified 668 proteins (23%) in the non-target set that have target-like properties. We suggest that drug discovery programmes would be more likely to succeed if new targets are chosen from this set or their homologues.
Authors: Alexander G Holman; Paul J Davis; Jeremy M Foster; Clotilde K S Carlow; Sanjay Kumar Journal: BMC Microbiol Date: 2009-11-28 Impact factor: 3.605