A machine learning approach for genome-wide prediction of morbid and druggable human genes based on systems-level data

Table 3 Statistical comparison of performances of classifiers trained on normal and without-one-feature druggability datasets

Missing feature ¹	Median AUC [min,max]²	N	W	W_c (two-tailed p = 0.05)³
ppi	0.819 [0.798,0.835]	10	27	8
metin	0.817 [0.803,0.834]	10	26	8
metout	0.817 [0.801,0.832]	9	20	6
regin	0.818 [0.799,0.83]	9	18	6
regout	0.818 [0.801,0.833]	10	26	8
c	0.821 [0.799,0.836]	10	21	8
identicalness	0.819 [0.8,0.836]	10	27	8
cent	0.814 [0.797,0.832]	10	18	8
inbet	0.821 [0.804,0.837]	10	25	8
inbetppi	0.819 [0.803,0.833]	10	25	8
inbetmet	0.82 [0.791,0.833]	10	26	8
inbetreg	0.818 [0.802,0.83]	9	19	6
numtissuesexp ⁴	0.806 [0.795,0.832]	9	11	6
avegexptec ⁵	0.814 [0.799,0.835]	10	23	8
Unknown	0.816 [0.796,0.832]	9	12	6
Cytoplasm	0.814 [0.794,0.834]	10	20	8
Endoplasmic reticulum	0.820 [0.799,0.834]	10	27	8
Mitochondrion	0.820 [0.796,0.831]	9	22	6
Nucleus	0.816 [0.793,0.831]	10	20	8
Other localization	0.821 [0.802,0.837]	9	20	6
Cellular component	0.82 [0.801,0.835]	10	25	8
Extracellular space	0.817 [0.8,0.837]	10	26	8
Golgi apparatus	0.812 [0.8,0.834]	10	24	8
Plasma membrane	0.781 [0.762,0.816]	10	1	8*
Median AUC [min,max] for normal datasets : 0.820 [0.801,0.835]

¹ See “Methods” and Additional file 1 for a description of features
² Of 10 datasets
³ According to table of critical values for W in [6]
⁴ The number of tissues (out of 32) in which the gene is expressed at least 5 transcripts per million (tpm) according to Reverter et al. [33]
⁵ The average expression in tpm among all the tissues in which the gene is expressed according to Reverter et al. [33]
* Difference statistically significant

ISSN: 1471-2164