Benchmarking of protein descriptor sets in proteochemometric modeling (part 1): comparative study of 13 amino acid descriptor sets

Table 2 Principal components resulting from the AAindex selection

Amino acid	PC1	PC2	PC3	PC4	PC5	PC6	PC7	PC8	Feature
Variance explained	0.43	0.24	0.08	0.06	0.04	0.03	0.03	0.02	n/a
Total variance explained	0.43	0.67	0.75	0.81	0.85	0.88	0.90	0.92	n/a
G	−5.70	−8.72	4.18	−1.35	−0.31	2.91	0.32	−0.11	−176196525
A	−0.10	−4.94	−2.13	1.70	−0.39	1.06	−1.39	0.97	1169372512
C	4.62	−3.54	1.50	−1.26	3.27	−0.34	−0.47	−0.23	892384356
V	5.04	−2.90	−2.29	1.38	0.06	0.08	1.79	−0.38	−58134849
L	5.76	−1.33	−1.71	0.63	−1.70	0.71	−0.05	−0.51	−590269326
I	6.58	−1.73	−2.49	1.09	−0.34	−0.28	1.97	−0.92	−1784790725
M	5.11	0.19	−1.02	0.15	0.13	−0.30	−2.95	0.50	−188476976
F	6.76	0.88	0.89	−1.12	−0.49	−0.55	−0.87	1.05	−1561345091
W	7.33	4.55	2.77	−2.41	−1.08	1.04	0.23	0.59	−816166777
Y	3.14	3.59	2.45	−1.27	−0.06	−0.29	1.99	0.30	1237879003
H	0.17	2.14	1.20	0.71	1.16	−0.38	−1.85	−2.79	−1970548995
T	−2.00	−1.77	−0.70	1.02	1.06	−1.20	0.74	1.65	−266397547
P	−3.82	−2.31	3.45	1.00	−3.22	−3.54	−0.36	−0.30	−576206913
S	−4.57	−2.55	−0.67	1.11	0.99	−1.02	0.11	0.65	−1481898440
D	−6.61	0.94	−3.04	−4.58	0.48	−1.31	0.10	0.94	1957532765
N	−4.88	0.81	0.14	−0.14	1.23	−0.65	1.02	−1.94	−1593568836
E	−5.10	2.20	−3.59	−2.26	−2.14	1.35	−0.45	−1.31	558044215
Q	−3.95	2.88	−0.83	0.52	0.90	0.55	−0.08	0.64	−1986194934
K	−4.99	5.00	0.70	3.00	−1.23	1.41	0.19	0.87	268201585
R	−2.79	6.60	1.21	2.07	1.67	0.76	0.00	0.32	1636879004

Shown are all eight principal components and the variance explained by these principal components. In addition, the features obtained from the hashing of the AAindex selection are shown. This column represents the feature based ProtFP. Not available is abbreviated by n/a.

ISSN: 1758-2946