A scalable machine-learning approach to recognize chemical names within large text databases

BMC Bioinformatics

Table 2 8-SPT as an example of how variation in chemical nomenclature affects information retrieval. Here, Ovid was used to map 5 variations of the chemical name 8-SPT to subject headings. Only adenosine and theophylline were common to each variant tested. 8-SPT is an adenosine receptor antagonist. For comparison, PubMed retrieved 341 unique records when using each of the keywords separated by "OR" in the query.

Chemical Name mapped to term(s)	8-p-sulphophenyltheophylline	8-sulphophenyltheophylline	8-p-sulfophenyltheophylline	8-SPT	8SPT
Adenosine	X	X	X	X	X
Adenosine Triphosphate		X		X	X
Aorta		X
Autonomic Nervous System	X
Cerebellum			X
Cerebral Cortex	X
Coronary Circulation			X
Coronary Vessels				X
Creatine Kinase					X
Endothelium, Vascular		X
Heart				X
Hippocampus					X
Hyperhomocysteinemia				X
Iris	X
Muscle Contraction	X
Muscle, Smooth			X
Myocardial Reperfusion Injury					X
Neurons					X
Neutrophils			X
Parasympathetic Nervous System	X
Phenethylamines		X
Rats, Wistar				X
Receptors, Adrenergic, alpha-1	X
Receptors, Cell Surface			X
Receptors, Purinergic	X		X	X
Receptors, Purinergic P1		X	X	X
Receptors, Purinergic P2		X
Spinal Cord		X
Synaptic Transmission	X
Theophylline	X	X	X	X	X
Vasodilation		X		X
Xanthines			X
# of documents retrieved via PubMed	22	251	47	297	15

ISSN: 1471-2105