Table 1.

Detected and corrected problems in the BioMeta database

Type of Problem	# in KEGG	# in BioMeta	# Corrected
Structure missing	1239	1106	133
Valence violation(s)	76	0	76
Incorrect constitution	unknown	unknown	107
Total (constitution)	1315	1106	316
Undefined stereo double bond(s)	35	32	3
Invalid sp3 stereocenter(s)	70	47	23
Ambiguous sp3 stereocenter(s)	46	0	46
Undefined sp3 stereocenter(s)	1398	865	533
Unspecified enantiomer	2326	1840	486
Undefined sp3 stereochemistry	554	366	188
Incorrect stereochemistry	unknown	unknown	69
Total (stereochemistry)	3990	2907	1152
Total corrected			1468

The table shows the validation and correction results of 12,815 entries present in both the KEGG Compound (version of October 25, 2005) and BioMeta databases. Note that the absence of a structure does not need to be an error – it may be a generic compound such as "acceptor" or "phosphorylated protein". Likewise, not all "unspecified enantiomer" cases need to be errors – a number of drugs may be racemic compounds. The row "total (stereochemistry)" is not the sum of the preceding cases because compounds may have multiple problems. The rows with the totals do not add up because of the "unknown" entries – if these numbers were known then the numbers would add up.