Click here to start

Table of contents

Use of Chemical Similarity in Drug Discovery

Basic principle of medicinal chemistry

Calculating Chemical Similarity

References to Daylight Fingerprints and Tanimoto index

Basic science behind Daylight fingerprints

Two important parameters to set ‘properly’

dbclus clustering algorithm

History of development of dbclus

Basic principle of dbclus algorithm: ‘many Merlin searches in an automated manner with Tanimoto similarity index controlling clustering algorithm’

Key steps in dbclus

dbclus output

Order independent dbclus algorithm

Final points about dbclus

Rule of thumb with Daylight fingerprints and Tanimoto

Slide with clusters at 0.9

Daylight fingerprints default settings

Two key patterns to look for when calibrating Daylight fingerprints


Design of representative sets

Designing sets for very low throughput experimental screens

Design of training/test sets for QSAR work

Data pre-processing before QSAR work

Example 2

QC application

HTS – false positives/negatives

Danger of ‘chemotype’ terminology

Influence of R1 and R2

Comparison of ‘goodness’ of different chemotypes

find_analogue code

In conclusion