Multiple regression analysis is a basic statistical tool used for QSAR studies in drug design. However, there is a risk or arriving at fortuitous correlations when too many variables are screened relative to the number of available observations. In this regard, a critical distinction must be made between the number of variables screened for possible correlation and the number which actually appear in the regression equation. Using a modified Fortran stepwise multiple-regression analysis program, simulated QSAR studies employing random numbers were run for many different combinations of screened variables and observations. Under certain conditions, a substantial incidence of correlations with high r2 values were found, although the overall degree of chance correlation noted was less than that reported in a previous study. Analysis of the results has provided a basis for making judgements concerning the level of risk of encountering chance correlations for a wide range of combinations of observations and screened variables in QSAR studies using multiple-regression analysis. For illustrative purposes, some examples involving published QSAR studies have been considered and the reported correlations shown to be less significant than originally presented through the influence of unrecognized chance factors.

No abstract

No abstract

A procedure is described in which an initial small group of compounds is selected, tested, and ordered according to potency. The potency order in the group is then compared to the tabulated potency order calculated for various parameter dependencies relating to hydrophobic, electronic, and steric effects. From this activity pattern analysis the probable operative parameters can be deduced and a new substituent selection made for the synthesis of potentially more potent analogues. Application of the method is illustrated with a series of examples. It differs from a previously described decision tree, single compound stepwise approach in that it involves the batchwise analysis of small groups of compounds, usually the preferred procedure for logistical reasons if the compounds are relatively easy to synthesize.

The quantitative structure-bioavailability relationship of 232 structurally diverse drugs was studied to evaluate the feasibility of constructing a predictive model for the human oral bioavailability of prospective new medicinal agents. The oral bioavailability determined in human adults was assigned one of four ratings and analyzed in relation to physicochemical and structural factors by the ORMUCS (ordered multicategorical classification method using the simplex technique) method. A systematic examination of various physicochemical parameters relating primarily to absorption, and structural elements which could influence metabolism, was carried out to analyze their effects on the bioavailabilty classification of drugs in the data set. Lipophilicity, expressed as the distribution coefficient at pH 6.5, was found to be a significant factor influencing bioavailability. The observation that acids generally had better bioavailability characteristics than bases, with neutral compounds between, led to the formulation of a new parameter, Delta log D (log D(6.5) - log D(7.4)), which proved to be an important contributor in improving the classification results. The addition of 15 structural descriptors relating primarily to well-known metabolic processes yielded a satisfactory QSAR equation which had a correct classification rate of 71% (97% within one class) and a Spearman rank correlation coefficient (R(s)) of 0.851, despite the diversity of structure and pharmacological activity in the compound set. In leave-one-out tests, an average of 67% of drugs were correctly classified (96% within one class) with an R(s) of 0.812. The relationship formulated identified significant factors influencing bioavailability and assigned them quantitative values expressing their contribution. The predictive power of the model was evaluated using a separate test set of 40 compounds, of which 60% (95% within one class) were correctly classified. Since the necessary physicochemical parameters can be calculated or estimated and the structural descriptors are obtained from an inspection of the structure, the model enables a rough estimate to be made of the prospective human oral bioavailability of unsynthesized compounds. Also, the model has the advantage of transparency in that it indicates which factors may affect bioavailabilty and the extent of that effect. This could be useful in designing compounds which are more bioavailable. Refinement of the model is possible as more bioavailability data becomes available. Potential uses are in drug design, prioritization of compounds for synthesis, and selection for detailed studies of early compound leads in drug discovery programs.