Table 2.

Compounds negatively correlated with T-ALL gene expression signature (P < .001)

Rank	Compound	Target/function	No. of instances	Mean connectivity score	Specificity
1	Alvespimycin	HSP90 inhibitor	12	−0.456	0.000
2	Wortmannin	PI3K inhibitor	18	−0.436	0.013
3	Vorinostat	HDAC inhibitor	12	−0.344	0.248
4	Geldanamycin	HSP90 inhibitor	15	−0.378	0.039
5	Trifluoperazine	Phenothiazine	16	−0.325	0.048
6	LY-294002	PI3K inhibitor	61	−0.437	0.012
7	Thioridazine	Phenothiazine	20	−0.423	0.065
8	Sirolimus	mTOR inhibitor	44	−0.339	0.048
9	Monorden	HSP90 inhibitor	22	−0.288	0.000
10	Trichostatin A	HDAC inhibitor	182	−0.299	0.198
11	Tanespimycin	HSP90 inhibitor	62	−0.349	0.076
12	Fluphenazine	Phenothiazine	18	−0.331	0.017
13	Puromycin	Antibiotic	4	−0.535	0.037
14	Prochlorperazine	Phenothiazine	16	−0.322	0.047
15	Cephaeline	Protein synthesis inhibitor	5	−0.384	0.066
16	Anisomycin	MAPK/JNK inhibitor	4	−0.462	0.042
17	Tonzonium bromide	Surfactant	4	−0.460	0.000
18	15-δ prostaglandin J2	PPAR-γ agonist	15	−0.283	0.105
19	Metixene	Anticholinergic	4	−0.393	0.000
20	Nortriptyline	Tricyclic antidepressant	4	−0.439	0.007
21	Emetine	Protein synthesis inhibitor	4	−0.466	0.053
22	Desipramine	Tricyclic antidepressant	4	−0.401	0.000
23	Terfenadine	Antihistamine	3	−0.534	0.028

Enrichment of both the up- and down-regulated genes from a given signature in the profiles of each treatment instance was estimated with a metric based on the Kolmogorov-Smirnov statistic and combined to produce a “connectivity score.” Instances were ranked by the descending order of connectivity scores; the frequency of an event can be taken as a P value. Specificity is defined as the frequency at which the enrichment of a set of instances in the ordered list of all instances in a given result is equaled or exceeded by the enrichment of that same set of instances in the set of results produced from queries executed with published, experimentally derived signatures.