Predicting subcellular location of protein with evolution information and sequence-based deep learning

BMC Bioinformatics

Table.2 Number of protein sequences in benchmark dataset 4802 created by Wei et al.

Subcellular location	Number of protein sequences
Nucleus	1720
Cytoplasm	1050
Plasma membrane	836
Extracellular	487
Mitochondria	407
Endosome	342
Golgi apparatus	272
Nucleolus	268
Lysosomes	125
Endoplasmic reticulum	120
Cytoskeleton	89
Centrosome	81
Peroxisome	67
Early endosomes	52
Nuclear envelope	47
Cytoplasmic vesicles	46
Basolateral plasma membrane	29
Synaptic vesicles	28
Microtubule	26
Apical plasma membrane	16
Late endosomes	16
Golgi trans face	11
Secretory granule	10
Tight junction	9
Golgi cis cisterna	7
Medial-golgi	7
Melanosome	6
Secretory vesicles	5
Cellular component	4
ERGIC	4
Inner mitochondrial membrane	4
Transport vesicle	4
Golgi trans cisterna	3
Total	6198

In this dataset, 4802 protein sequences are identified in 33 subcellular locations. The first column is the name of the subcellular covered by this dataset, and the second column is the number of proteins located at each subcellular location. The total number of subcellular locations is 6198 since each sequence can be found in multiple subcellular locations. The sequences distribute at those 33 locations unevenly. 35.8% of sequence samples are located at Nucleus and 21.9% sequences are located in Cytoplasm, while only 3 sequences are identified in Golgi Trans Cisterna. The number of positive cases in this dataset is 6198, and the positive case rate is 3.9% (6198/(4802 × 33))

ISSN: 1471-2105