The invention relates to a process for drug candidate identification, said process comprising the steps of: (1) obtaining a computerised representation of the three-dimensional structure of a binding site on the surface of a biological macromolecule; (2) generating a computerised model of the functional structure of said binding site which may be used to identify favourable and unfavourable interactions between the binding site and a drug candidate molecule; (3) identifying a molecular fragment (or "template" T) capable of placement within said binding site and capable of carrying at least one (preferably a plurality (ie. at least two) and especially preferably at least 3) substituent group, said molecular fragment either being capable of being synthesized from reagent compounds accessible in substituted form whereby to import said substituent groups on synthesis of said molecular fragment or being present in an accessible reagent compound capable of substitution with said substituent groups by reaction with further accessible reagent compounds; ; (4) generating a set of lists of accessible reagent compounds (eg. a1-A, a2-A, a3-A, etc, b1-B, b2-B, b3-B, etc, c1-C, c2-C, c3-C, etc), the lists being such that a combination of compounds taken from each list (eg. a1-A, b3-B and c11-C) may be reacted to produce a candidate compound comprising said molecular fragment carrying a plurality of substituent groups (eg. a1b3 c11T) thereby generating a first virtual library of candidate compounds being the theoretical set of compounds producible by reaction of the members of said lists (ie. a1b1c1T, a1b1c2T, a1b2c1 T etc), each member of each list comprising a component (eg. A,B,C, etc.) common to the other members of that list and a component (eg. a1, b1, c1, etc) unique within that list; (5) for each said list limiting the number of members thereof using a first set of exclusion rules thereby to generate a restricted second virtual library of candidate compounds, the operation of said first set of rules involving for each member of each list computerised comparison for favourable or unfavourable interactions between said computerised model and a structure comprising said molecular fragment and a substituent deriving from the unique component within said list of that member, the molecular fragment and the computerized model being held in fixed spatial relationship to each other for said comparison; (6) evaluating and ranking by computer the members of said second virtual library for favourable and unfavourable interactions with said computerised model and thereby generating a restricted third virtual library of candidate compounds ranked as having favourable interactions; (7) optionally, selecting from said third virtual library at least one further molecular fragment and repeating steps (4), (5) and (6) to generate an alternative third virtual library; (8) screening said third virtual library using a second set of exclusion rules thereby to generate a restricted fourth virtual library of candidate compounds comprising compounds which are candidates for synthesis and experimental evaluation for drug efficacy; (9) synthesizing some or all candidate compounds of said fourth virtual library to produce a candidate compound library; (10) experimentally evaluating the compounds of said candidate compound library for drug efficacy; (11) analysing the experimental efficacy data generated in step (10) for structure-activity relationship information; (12) using the information derived in step (11) selecting a revised set of lists of accessible reagent compounds, said lists being expanded to include selected reagents not present in the restricted lists generated in step (5) and optionally restricted to exclude selected reagents present in the restricted lists generated in step (5); (13) repeating steps (6) and (7) to identify further compounds which are candidates for synthesis and experimental evaluation for drug efficacy; (14) synthesising and experimentally evaluating said further compounds for drug efficacy; (15) if required repeating steps (11) to (14) one or more times; ; (16) identifying as a lead candidate a compound synthesized and experimentally evaluated as above. The process of the invention is characterised by the rapid generation of a relatively small set of readily synthesisable candidate compounds with a high success rate in terms of drug efficacy and hence a high predictive value for directing subsequent iterations.