Table 4.

Results (Top 1, 5 and 10 ranks) using PubChem formula queries on three additional datasets

Weight term	Score Term	Weights
${\mathit{\omega }}_{\mathrm{Frag}}$	$S_{C_{\mathrm{Frag}}}$	1	1	1	0	1	0	0
${\mathit{\omega }}_{\mathrm{RTs}}$	$S_{C_{\mathrm{RT}}}$	1	1	0	1	0	1	0
${\mathit{\omega }}_{\mathrm{Refs}}$	$S_{C_{\mathrm{Refs}}}$	1	0	1	1	0	0	1

Dataset	Metric	Ranks
UF (n = 225)	Top 1 ranks	164 (73 %)	9	163	159	3	2	157
UF (n = 225)	Top 5 ranks	186 (83 %)	48	189	189	36	13	199
UF (n = 225)	Top 10 ranks	191 (53 %)	77	196	192	61	25	204
EQex (n = 289)	Top 1 ranks	235 (81 %)	33	232	230	26	11	223
EQex (n = 289)	Top 5 ranks	263 (91 %)	87	260	258	88	38	276
EQex (n = 289)	Top 10 ranks	270 (93 %)	132	269	263	139	55	280
EQexPlus (n = 310)	Top 1 ranks	190 (61 %)	32	183	182	21	8	181
EQexPlus (n = 310)	Top 5 ranks	238 (77 %)	84	246	238	83	28	243
EQexPlus (n = 310)	Top 10 ranks	254 (82 %)	115	258	247	121	37	256

The weights indicate where ranking parameters were included (1) or excluded (0) from the candidate ranking. Retention time score calculation was performed using the XLOGP3 values of PubChem. ${\mathit{\omega }}_{\mathrm{Refs}}·S_{C_{\mathrm{Refs}}}={\mathit{\omega }}_{\mathrm{Refs}}·S_{C_{\mathrm{PNP}+\mathrm{PPC}}}$. See text for explanations