From: WIECZOREK Samuel 175264 <WIECZOREK@ds...>  20030912 06:50:41

Hi everyone, I need an explanation about the generation of molecular fingerprints in = CDK. In the classe Fingerprinter, I understood this : a complete liste of pathes (or substructure) is generated up to the = limit given by the attribute serachDepth. Each of these elements is then = hashcoded with the function hashCode. The corresponding hashCode is used as a = seed for generating a pseudo random number which is a position in the = fingerprint. Then, you set only the bit at this position to 1. So, one bit represent = one hashcoded path. I have seen a difference with the fingerprints described in Daylight = System. There, I understood that the entire hashcode (generated by a path) is = added to the finger print at a random position. It is not only one bit that = is set to 1.=20 I would like to know which version is correct. In fact, I don't see any = big difference between the implemantation of fingerprints in CDK and the structural keys where one bit represent the presence/abscence of one = path (or pattern). The only difference in CDK is that the position of the = bit is not known before. Thanks for your help Best regards Samuel Wieczorek=20 Samuel Wieczorek Service de Developpements pour la Bioinformatique Direction des Sciences du Vivant Commissariat =E0 l'Energie Atomique T=E9l. : 04.38.78.98.23 FAX. : 04.38.78.43.37 