Community Research and Development Information Service - CORDIS


By means of clustering according to structure similarity, large databases containing several tens of thousands of chemicals can easily be managed. A new methodology is proposed in which statistical criteria are used to decide the optimal threshold in the clustering process. All the clusters in the training set showed correlation of logWS with logK(ow) and melting point, except one. The latter covers all chemicals with melting points below room temperature, resulting in a logWS-logK(ow) relationship. This approach resulted in 3 QSARs with reasonably good predictive capabilities. The models resulting from the smaller clusters are characterised by high correlation coefficients, describing the cluster itself very well but performing close to random models.

Additional information

Authors: NOUWEN J, JRC Ispra (IT);HANSEN B, JRC Ispra (IT)
Bibliographic Reference: Article: Quantitative Structure-Activity Relationships (QSAR) (1995)
Record Number: 199511219 / Last updated on: 1995-10-10
Original language: en
Available languages: en