Using complexity measure factor to predict protein subcellular location

Abstract

Recent advances in large-scale genome sequencing have led to the rapid accumulation of amino acid sequences of proteins whose functions are unknown. Because the functions of these proteins are closely correlated with their subcellular localizations, it is vitally important to develop an automated method as a high-throughput tool to timely identify their subcellular location. Based on the concept of the pseudo amino acid composition by which a considerable amount of sequence-order effects can be incorporated into a set of discrete numbers (Chou, K. C., Proteins: Structure, Function, and Genetics, 2001, 43: 246-255), the complexity measure approach is introduced. The advantage by incorporating the complexity measure factor as one of the pseudo amino acid components for a protein is that it can more effectively reflect its overall sequence-order feature than the conventional correlation factors. With such a formulation frame to represent the samples of protein sequences, the covariant-discriminant predictor (Chou, K. C. and Elrod, D. W., Protein Engineering, 1999, 12: 107-118) was adopted to conduct prediction. High success rates were obtained by both the jackknife cross-validation test and independent dataset test, suggesting that introduction of the concept of the complexity measure into prediction of protein subcellular location is quite promising, and might also hold a great potential as a useful vehicle for the other areas of molecular biology.

Boosting classifier for predicting protein domain structural class

“…Moreover, it has not escaped our notice that the LogitBoost classifier can also be used to predict other protein attributes, such as subcellular localization [35][36][37][38][39][40], membrane types [41][42][43][44][45], enzyme family and subfamily classes [46][47][48][49], enzyme active sites [50,51], G-protein coupled receptor classification [52,53], and protein quaternary structure types [54], among many others. Table 2 of Zhou [18].…”

High performance set of PseAAC and sequence based descriptors for protein classification

“…The first class comprises the Chou's pseudo amino acid (PseAA) composition , probably the most used feature extractor for proteins, and its variants. To avoid losing important information hidden in protein sequences, the pseudo amino acid composition (PseAAC) was proposed (Chou, 2001;Chou, 2005) to replace the simple amino acid composition (AAC) for representing the sample of a protein.…”

“…For example, cellular 4 automata image Xiao et al, 2008a;Xiao et al, 2009a;Xiao et al, 2006a), complexity measure factor (Xiao et al, 2006b;Xiao et al, 2005); Grey dynamic model Xiao et al, 2008b); functional domain composition (Xiao et al, 2009b).…”

A New Ensemble Scheme for Predicting Human Proteins Subcellular Locations

“…Early attempts were based on the decision of a single learner. Covariant Discriminant Classifier (CDC) was attempted using different feature extraction techniques [5][6][7][8][9][10]. Support Vector Machines (SVM) classifier was tried with Functional Domain Composition [11] features.…”