Supplemental Tables for “Substring Selection for Biomedical Document Classification”

Supplement Table 1. Summary of the five PTM datasets

PTM types	Positives	Negatives
Acetylation	55	868
Glycosylation	41	711
Methylation	27	171
Phosphorylation	79	389
Hydroxylation	27	133

Supplement Table 2. The Unlabeled Glycosylation Abstracts

Attributes Used for Ranking	Stemmed Word	Substring
Full Dataset	Top50	Top50
	Bottom50	Bottom50