1 Computational Approaches(1/7)  Computational methods can be divided into four categories: prediction methods based on  (i) The overall protein amino acid composition  (ii) Known targeting sequences  (iii) Sequence homology and/or motifs  (iv) A combination of several sources of information (hybrid methods)

2 Computational Approaches(2/7) (i) The overall protein amino acid composition(1/2)  Nakashima and Nishikawa  Method for discriminating between intracellular and extracellular proteins  Using the distance between the overall amino acid composition vectors  Cedano et al  ProtLock for predicting five classes of subcellular localizations  Extracellular, intracellular, integral membrane, anchored membrane, and nuclear  Reinhardt and Hubbard  NNPSL, an approach using artificial neural networks (ANNs)  Predicting four eukaryotic and three prokaryotic subcellular localizations  Chou et al  SVM-based method for predicting twelve different subcellular localization taking sequence order effects into account

3 Computational Approaches(3/7) (i) The overall protein amino acid composition(2/2)  Huang et al  Using Fuzzy k-NNs algorithm  Describe the dipeptide composition of the whole protein sequence for eleven different localizations  Yu, C.S. (CELLO method)  Prediction of five subcellular localizations in Gram-negative bacteria  Based on the composition of peptides of varying lengths  Andrade et al  First to incorporate structural information into the amino acid composition vectors  Composition of eukaryotic proteins with known structure was used  The rationale behind this approach  The interiors of proteins have stayed fairly constant during evolution

4 Computational Approaches(4/7) (ii) Known targeting sequences  Gunnar von Heijne (TargetP)  The most comprehensive method based on N-terminal targeting sequences  Prediction of chloroplast, mitochondrial, secretory pathway, and other proteins  Claros M.G. (MitoProt and Predotar)  Specifically discriminate chloroplast from mitochondrial proteins  Bannai H. et al (iPsort)  Using knowledge-based rules for prediction based on protein sequence features

5 Computational Approaches(5/7) (iii) Sequence homology and/or motifs  Marcotte et al  Assigns the subcellular localization by constructing phylogenetic profiles of the proteins  Cokol M et al(PredictNLS)  Specialized on recognizing nuclear proteins  Based on a collection of nuclear localization sequences  Lu et al(Proteome Annalyst)  Based on SWISS-PROT keywords and the annotation of homologous proteins

6 Computational Approaches(6/7) (iv) Hybrid methods  Nakai K and Kanehisa(PSORT)  One of the first methods developed for predicting the subcellular localization  Using the overall amino acid composition, N-terminal targeting sequence information, and motifs  This method uses a set of knowledge-based "if-then" rules  Predicts 14 animal and 17 plant subcellular localizations  PSORT II and and PSORT-B-Extensions of the PSORT  Drawid and G  Method that incorporates information about sequence motifs, overall sequence properties and mRNA expression levels  Based on a Bayesian prediction model and was tested on the yeast genome  Guda C et al(MITOPRED)  Specialized for predicting mitochondrial proteins  Based on amino acid composition