Table 1

Selected disease classes and their associated classification performance.

UMLS concepts

Datasets

Phenotype distance score

(p-value)

ManiSVM accuracy

SVM accuracy

C0027651 (Neoplasms), C0027660 (Neoplasms, Glandular and Epithelial), C0040300 (Body tissue), C0007097 (Carcinoma), C0027653 (Neoplasms by Site), C0027652 (Neoplasms by Histologic Type)

GDS1070

GDS1321

GDS1479

GDS505

3.50E-05

0.8421

0.6018

C0018981 (Hemic and Lymphatic Diseases), C0005773 (Blood Cells), C0018939 (Hematological Disease)

GDS1257

GDS1392

GDS539

GDS1320

GDS390

7.10E-05

0.8047

0.6253

C0007682 (CNS disorder), C0006111 (Brain Diseases), C0027765 (nervous system disorder)

GDS1331

GDS1726

GDS1065

9.99E-03

0.7569

0.6483

C0021311 (Infection), C0004615 (Bacterial Infections and Mycoses)

GDS1428

GDS1022

GDS539

GDS711

GDS1726

GDS1397

2.36E-04

0.7498

0.5253

Liu et al. BMC Bioinformatics 2009 10(Suppl 1):S25   doi:10.1186/1471-2105-10-S1-S25