Model fitting for small skin permeability data sets: hyperparameter optimisation in Gaussian Process Regression

Forecasting and Optimizing Dual Media Filter Performance via Machine Learning

Four different machine learning algorithms, including Decision Tree (DT), Random Forest (RF), Multivariable Linear Regression (MLR), Support Vector Regressions (SVR), and Gaussian Process Regressions (GPR), were applied to predict the performance of a multi-media filter operating as a function of raw water quality and plant operating variables. The models were trained using data collected over a seven year period covering water quality and operating variables, including true colour, turbidity, plant flow, and chemical dose for chlorine, KMnO₄, FeCl₃, and Cationic Polymer (PolyDADMAC). The machine learning algorithms have shown that the best prediction is at a 1-day time lag between input variables and unit filter run volume (UFRV). Furthermore, the RF algorithm with grid search using the input metrics mentioned above with a 1-day time lag has provided the highest reliability in predicting UFRV with a RMSE and R² of 31.58 and 0.98, respectively. Similarly, RF with grid search has shown the shortest training time, prediction accuracy, and forecasting events using a ROC-AUC curve analysis (AUC over 0.8) in extreme wet weather events. Therefore, Random Forest with grid search and a 1-day time lag is an effective and robust machine learning algorithm that can predict the filter performance to aid water treatment operators in their decision makings by providing real-time warning of the potential turbidity breakthrough from the filters.

The absorption of polycyclic aromatic hydrocarbons into the skin to elicit cutaneous inflammation: The establishment of structure-permeation and in silico-in vitro-in vivo relationships

Polycyclic aromatic hydrocarbons (PAHs) can induce skin toxicity. Although some investigations have been conducted to assess the skin toxicity of different PAHs, few comparisons using a series of PAHs with different ring numbers and arrangements have been done. We aimed to explore the skin absorption of 6 PAH compounds and their effect on cutaneous inflammation. In vitro skin permeation was rated by Franz cell with pig skin. Molecular docking was employed to compute the PAH interaction with stratum corneum (SC) lipids. Cultured keratinocytes were exposed to PAHs for analyzing cytotoxicity, cyclooxygenase (COX)-2, prostaglandin E₂ (PGE₂), chemokines, and differentiation proteins. The in vivo topical PAH exposure in mice was characterized by skin absorption, transepidermal water loss (TEWL), PGE₂ level, and histology. The skin deposition from the aqueous vehicle increased following the increase of PAH lipophilicity and molecular size, with benzo[a]pyrene (5-ring PAH) showing the greatest absorption. Pyrene was the compound showing the highest penetration across the skin (flux). Although the PAHs fluoranthene, pyrene, chrysene, and 1,2-benzanthracene all had 4 rings, the skin permeation was quite different. 1,2-Benzanthracene showed the greatest absorption among the 4-ring compounds. The PAHs with higher absorption exhibited stronger interaction with SC lipids according to the in silico modeling. Chrysene and 1,2-benzanthracene generally showed the highest COX-2 and PGE₂ expression, followed by benzo[a]pyrene. The lowest COX-2 and PGE₂ upregulation was observed for naphthalene (2-ring PAH). A contrary tendency was detected for the upregulation of chemokines. Filaggrin and integrin β1 in keratinocytes were suppressed at a comparable level by all PAHs. The skin's absorption of PAHs showed strong in vivo-in vitro correlation. 1,2-Benzanthracene and benzo[a]pyrene highly disrupted the skin barrier and elevated the inflammation in vivo. The tendency toward in vivo inflammation caused by various PAHs could be well predicted by the combined estimation using in vitro skin absorption and a keratinocyte bioassay. This study also established the structure-permeation relationship (SPR) of PAHs.

Development of a Gaussian Process - feature selection model to characterise (poly)dimethylsiloxane (Silastic® ) membrane permeation

OBJECTIVES

The current study aims to determine the effect of physicochemical descriptor selection on models of polydimethylsiloxane permeation.

METHODS

A total of 2942 descriptors were calculated for a data set of 77 chemicals. Data were processed to remove redundancy, single values, imbalanced and highly correlated data, yielding 1363 relevant descriptors. For four independent test sets, feature selection methods were applied and modelled via a variety of Machine Learning methods.

KEY FINDINGS

Two sets of molecular descriptors which can provide improved predictions, compared to existing models, have been identified. Best permeation predictions were found with Gaussian Process methods. The molecular descriptors describe lipophilicity, partial charge and hydrogen bonding as key determinants of PDMS permeation.

CONCLUSIONS

This study highlights important considerations in the development of relevant models and in the construction and use of the data sets used in such studies, particularly that highly correlated descriptors should be removed from data sets. Predictive models are improved by the methodology adopted in this study, notably the systematic evaluation of descriptors, rather than simply using any and all available descriptors, often based empirically on in vitro experiments. Such findings also have clear relevance to a number of other fields.

Development of flavanone and its derivatives as topical agents against psoriasis: The prediction of therapeutic efficiency through skin permeation evaluation and cell-based assay

Flavonoids inhibit skin inflammation. Previous study suggests that the flavonoids with flavanone backbone were beneficial to penetrate into the skin. We aimed to investigate the possibility of psoriasis treatment by topically applied flavanone and its derivatives including naringenin, hesperetin, 6-hydroxyflavanone, flavanone, and 6-bromoflavone. The skin absorption of the compounds was determined by Franz cells. Molecular modeling was used to compute the physicochemical and molecular parameters of the penetrants in order to elucidate the correlation between structure and permeation. Among the compounds tested, flavanone showed the greatest skin absorption. The in vitro skin absorption predicted efficient skin targeting of 6-bromoflavone with minimal risk of circulation absorption. The permeation of naringenin was remarkably enhanced 13-fold in the barrier-defective skin mimicking inflamed skin. The penetrants with fewer hydrogen bond number, total polarity surface, and molecular volume were advantageous for facile skin absorption. In the cell-based study, IL-1β inhibition in imiquimod (IMQ)-stimulated keratinocytes was increased following the increase in compound lipophilicity. Naringenin, a flavanone analog with three hydroxyl moieties, could suppress IL-6 overexpression to baseline control. We assessed the anti-inflammatory potency of the chemicals in comparison with tacrolimus as reference in a psoriasis-like mouse model. Flavanone was found to mitigate scaling and epidermal hyperplasia at a higher level than naringenin. Flavanone lessened IL-6 overexpression by 80% in the psoriasiform plaque. The skin barrier function recorded by transepidermal water loss (TEWL) was recovered by naringenin but not flavanone. The experimental data indicate that naringenin and flavanone are potential candidates for anti-psoriatic therapy.

The influence of diffusion cell type and experimental temperature on machine learning models of skin permeability

OBJECTIVES

The aim of this study was to use Gaussian process regression (GPR) methods to quantify the effect of experimental temperature (T_exp ) and choice of diffusion cell on model quality and performance.

METHODS

Data were collated from the literature. Static and flow-through diffusion cell data were separated, and a series of GPR experiments was conducted. The effect of T_exp was assessed by comparing a range of datasets where T_exp either remained constant or was varied from 22 to 45 °C.

KEY FINDINGS

Using data from flow-through diffusion cells results in poor model performance. Data from static diffusion cells resulted in significantly greater performance. Inclusion of data from flow-through cell experiments reduces overall model quality. Consideration of T_exp improves model quality when the dataset used exhibits a wide range of experimental temperatures.

CONCLUSIONS

This study highlights the problem of collating literature data into datasets from which models are constructed without consideration of the nature of those data. In order to optimise model quality data from only static, Franz-type, experiments should be used to construct the model and T_exp should either be incorporated as a descriptor in the model if data are collated from a range of studies conducted at different temperatures.