A text-mining system for extracting metabolic reactions from full-text articles

BMC Bioinformatics

Table 1 Assignments of the entities enzyme (E), substrate (S) and product (P) for a sample sentence

L-Arabinose isomerase	L-arabinose	L-ribulose	n-arabinose
E	S	P	P
E	S	S	P
E	P	S	S
E	P	P	S
E	S	P
E	S		P
E		S	P
E	P	S
E	P		S
E		P	S

The ten assignments of E, S and P for the sentence “L-Arabinose isomerase catalyzes the conversion of L-arabinose to L-ribulose, the first step in the utilization of n-arabinose by Escherichia coli B/r”. Given that L-Arabinose isomerase is the only tagged protein, it is deemed to be the enzyme in all cases, whereas different numbers and orderings of substrates and products are possible, given the presence of three tagged small molecules (L-arabinose, L-ribulose and n-arabinose). Note that other potential orderings (namely E-P-S-P and E-S-P-S) are not considered, as they are deemed highly unlikely to occur in practice.

ISSN: 1471-2105