. Author manuscript; available in PMC: 2011 Aug 1.

Published in final edited form as: J Biomed Inform. 2010 Mar 27;43(4):510–519. doi: 10.1016/j.jbi.2010.03.008

Table 4. Precision and recall results as thresholds are applied.

Hypothesis Set Count – number of chemicals in hypothesis set, Found FL Chemicals – number of future linked chemicals found by our process, Found FL Articles – number of articles associated with the found future linked chemicals. Precision, Recall, and Article Recall are calculated from the hypothesis set when the protein count (protct) threshold is applied.

Threshold Applied	Hypothesis Set Count	Found FL Chemicals	Found FL Articles	Precision	Recall	Article Recall
none	4725	154	552	0.03	0.870	0.909
protct > 1	2658	138	529	0.05	0.780	0.871
protct > 2	1867	131	511	0.07	0.740	0.842
protct > 3	1454	123	498	0.08	0.695	0.820
protct > 4	1223	114	486	0.09	0.644	0.801
protct > 5	1034	105	460	0.10	0.593	0.758
protct > 6	888	93	424	0.10	0.525	0.699
protct > 7	801	89	412	0.11	0.503	0.679
protct > 8	739	86	406	0.12	0.486	0.669
protct > 9	674	86	406	0.13	0.486	0.669
protct > 10	617	82	399	0.13	0.463	0.657