Table 6.

Statistical Analysis of the Improvement in Rank

Search Size:	10	20	30	40	50	60	70	80	90
Therapuetic Use
analgesic	0.136	0.087	0.041	0.020	0.001	2.0E-04	0.004	0.001	0.002
anti-convulsant	0.811	0.529	0.932	0.636	0.918	0.856	0.478	0.481	0.512
anti-inflammatory	0.151	0.040	0.034	0.033	0.002	0.001	0.003	0.002	0.004
anti-neoplastic	0.412	0.842	0.943	0.928	0.577	0.804	0.919	0.909	0.779
anti-proliferative	0.148	0.553	0.778	0.659	0.987	0.881	0.942	0.837	0.837
anti-psychotic (neurole	0.905	0.852	0.686	0.674	0.529	0.129	0.063	0.024	0.006
cardiovascular agent	0.647	0.267	0.496	0.496	0.540	0.795	0.984	0.801	0.687
immunosuppressive agent	0.047	0.664	0.642	0.550	0.525	0.525	0.554	0.323	0.323
tranquilizer	1.000	0.580	0.579	0.921	0.665	0.619	0.413	0.290	0.159

A sensitivity analysis using multiple paired t-tests was performed to analyze the difference in the rank of true positives returned by the direct and indirect method at different search thresholds for rank k. Each cell contains the respective probability. All significant (P < .05) findings are confirmed to be in favor of the indirect method.