Table 1.

Features extracted.

Feature	Value
word	all words in the training data
orthography	capital, symbol, etc. (see Table 2)
prefix	1, 2, or 3 gram of the starting letters of a word
suffix	1, 2, or 3 gram of the ending letters of a word
part of speech	Brill tagger
preceding class	-2, -1
gene/protein name dictionary	protein names collected from SWISS-PROT and TrEMBL