A complete map of potential pathogenicity markers of avian influenza virus subtype H5 predicted from 11 expressed proteins

Table 1 The training data

Protein	H5N1		Non-H5N1		Total		All Features	Significant Features
Protein	HP	LP	HP	LP	HP	LP	All Features	Significant Features
HA	1377	54	48	512	1425	566	616	82
NA	551	32	23	264	574	296	593	114
M1	161	9	13	52	174	61	329	16
M2	186	9	14	63	200	72	98	18
NS1	425	16	22	148	447	164	249	71
NS2	202	3	14	53	216	56	129	25
NP	294	12	22	113	316	125	511	22
PA	465	22	25	235	490	257	730	57
PB1	405	26	25	223	430	249	775	44
PB2	446	26	23	247	469	273	783	62
PB1-F2	135	16	15	114	150	130	101	40

The HP and LP columns represent the number of highly pathogenic and low pathogenic sequences in each of the proteins, respectively. The ‘All features’ column is the total number of features (i.e. AA’s) from which significant features are selected with Monte Carlo Feature Selection

ISSN: 1471-2180