Biomedical subjects

S E Pack

Publications and source records attributed to S E Pack.

4 recordsLinked to original sources

PubMed1997

EVA: a new theoretically based molecular descriptor for use in QSAR/QSPR analysis.

A new descriptor of molecular structure, EVA, for use in the derivation of robustly predictive QSAR relationships is described. It is based on theoretically derived normal coordinate frequencies, and has been used extensively and successfully in proprietary chemical discovery programmes within Shell Research. As a result of informal dissemination of the methodology, it is now being used successfully in related areas such as pharmaceutical drug discovery. Much of the experimental data used in development remain proprietary, and are not available for publication. This paper describes the method and illustrates its application to the calculation of nonproprietary data, log P(ow), in both explanatory and predictive modes. It will be followed by other publications illustrating its application to a range of data derived from biological systems.

Drug Design↗

PubMed1997

Finite mixture models for proportions.

Six data sets recording fetal control mortality in mouse litters are presented. The data are clearly overdispersed, and a standard approach would be to describe the data by means of a beta-binomial model or to use quasi-likelihood methods. For five of the examples, we show that beta-binomial model provides a reasonable description but that the fit can be significantly improved by using a mixture of a beta-binomial model with a binomial distribution. This mixture provides two alternative solutions, in one of which the binomial component indicates a high probability of death but is selected infrequently; this accounts for outlying litters with high mortality. The influence of the outliers on the beta-binomial fits is also demonstrated. The location and nature of the two main maxima to the likelihood are investigated through profile log-likelihoods. Comparisons are made with the performance of finite mixtures of binomial distributions.

Animals↗

PubMed1989

QUAD: a computer package for the analysis of QUantal Assay Data.

The computer package QUAD has been developed at the University of Kent, U.K. It is menu driven and written in Advanced BASIC. It runs on IBM PC compatible machines equipped with a suitable graphics facility such as CGA or simulated CGA. QUAD is available on a floppy disk, for a small handling charge. QUAD has four main functions: it performs a logit analysis of quantal assay data; it provides a flexible way of analysing the data, allowing dose transformations and providing alternative confidence intervals for EDp values; it produces a range of diagnostics for assessing the fit of models to data; it provides and fits two families of extended models, each containing the logit as a special case. The package makes use of the latest statistical research, and fitted models are displayed by means of the good graphics facilities available on microcomputers. This document describes the facilities available in detail, and provides and discusses, illustrations of the package at work. QUAD has been designed as a pilot package. Further additions and developments are planned and described later.

Animals↗

PubMed1986

Hypothesis testing for proportions with overdispersion.

The properties of likelihood ratio tests and simpler t-tests are investigated by simulation under an assumed beta-binomial model for parameter values typically found in toxicological studies. It is found that likelihood ratio methods are at least as powerful as the simpler approaches and in certain situations can be significantly more powerful.

Analysis of Variance↗