Predicting Interactions between Virus and Host Proteins Using Repeat Patterns and Composition of Amino Acids

Table 5

Training (TR) and test (TS) datasets for assessing the applicability of the SVM model to new viruses and to new hosts. The average sequence similarity between proteins in TR and those in TS was analyzed using EMBOSS Needle tool [20].

Proteins in training datasets	Target proteins in test datasets	Average sequence similarity (%)
25 virus proteins in TR1	11 HCV proteins in TS1	5.03
	12 SARS virus proteins in TS2	5.20
	10 H1N1 virus proteins in TS3	5.03
	11 HPV-16 proteins in TS4	3.12
	46 HIV-1 proteins in TS5	3.56
522 human proteins in TR2	141 Mus musculus proteins in TS6	9.20
	87 Bos taurus proteins in TS7	9.07
	79 Rattus norvegicus proteins in TS8	9.76
	38 Sus scrofa proteins in TS9	8.70
	64 Escherichia coli K-12 proteins in TS10	8.04