The application of Gaussian processes in the prediction of percutaneous absorption

Abstract

A non-linear approach was more appropriate than QSPRs or SLNs for the analysis of the dataset employed herein, as the prediction and confidence values in the prediction given by the Gaussian process are better than with other methods examined. Gaussian process provides a novel way of analysing skin absorption data that is substantially more accurate, statistically robust and reflective of our empirical understanding of skin absorption than the QSPR methods so far applied to skin absorption.

Predicting drug absorption rates through human skin

“…The single layer network (NN), also a linear model, has done better and does improve on the baseline. The Gaussian Process model is clearly the best and this confirms our earlier results [5] that suggested that skin permeability prediction needs a nonlinear model. Figure 1 illustrates the non-linearity of this problem.…”

“…In recent years there has been increasing attention paid to the problem of predicting the rate at which a substance will pass through human skin [1][2][3][4][5][6]. Interest in this issue has been driven by the pharmaceutical industry, where it is important to be able to predict the rate at which a drug will pass through the skin and into the bloodstream.…”

“…There has also been a small amount of work looking at nonlinear models [2], although none of these have used The data set used here consists of the original Potts and Guy data supplemented by a variety of data from other sources. It contains data on 142 molecules and a detailed description can be found in [5]. As well as molecular weight and lipophilicity we also have the four following additional molecular features that are thought to be important in determining a molecules ability to pass through skin: So in summary our data set is 142 6-ary vectors each of which has a measured value of its skin permeability coefficient log K p .…”

A review of the current state of the art of physiologically-based tests for measuring human dermal in vitro bioavailability of polycyclic aromatic hydrocarbons (PAH) in soil

“…Machine learning algorithms are also applied in QSAR, as shown in recent studies: gene expression programming (Wang et al,, 2008); Gaussian profile (Moss et al,, 2009) and similar stochastic techniques by others (Table 2). Cronin and Schultz (2003) recognised that QSAR outputs should achieve the following criteria: 1) a well-defined and measurable endpoint; 2) use a chemically and biologically diverse dataset; 3) be based on chemical descriptors that are consistent with the endpoint; 4) use appropriate statistical methods; and 5) have a strong mechanistic basis.…”

Gaussian process regression with functional covariates and multivariate response

“…See for example the seminal book [3] and the references therein for details. In chemometrics and related areas, GPR has been applied to a range of problems, such as calibration of spectroscopic analysers [4,5], response surface modelling [6], system identification [7], ensemble learning [5,8], prediction of transmembrane pressure [9], and prediction of percutaneous absorption [10,11], among others.…”