Table 1.

The number of proteins in the original Huh et al. Dataset (2003) and three training datasets

Subcellular localization	Huh et al. Dataset	Dataset-I	Dataset-II	Dataset-III
1. Actin	32	32	27	27
2. Bud	25	25	19	19
3. Bud neck	61	61	48	48
4. Cell periphery	130	130	98	98
5. Cytoplasm	1782	1782	1472	1472
6. Early golgi	54	54	39	39
7. Endosome	46	46	37	37
8. ER	292	292	207	207
9. ER to golgi	6	6	5	5
10. Golgi	41	41	30	30
11. Late golgi	44	44	38	38
12. Lipid particle	23	23	15	15
13. Microtubule	20	20	17	17
14. Mitochondrion	522	522	389	389
15. Nuclear periphery	60	60	38	38
16. Nucleolus	164	164	122	122
17. Nucleus	1446	1446	1126	1126
18. Peroxisome	21	21	16	16
19. Punctate composite	137	137	91	91
20. Spindle pole	61	61	27	27
21. Vacuolar membrane	58	58	47	47
22. Vacuole	159	159	124	124
Total number of classified proteins, $\tilde{N}$	5184	5184	4032	4032
Total number of different proteins, N	3914	3914	3017	3017
Dimension of features		9620D	2372D	11992D
Coverage		100%	77.08% $\left(\frac{3017}{3914}\right)$	77.08% $\left(\frac{3017}{3914}\right)$