Evaluating point-prediction uncertainties in neural networks for protein-ligand binding prediction
Not Available
97 MATHEMATICS AND COMPUTING↗
Engineering topics
McLoughlin, Kevin S.
Publications and source records attributed to McLoughlin, Kevin S..
Model Choice Metrics to Optimize Profile-QSAR Performance
Predicting molecular activity against protein targets is difficult because of the paucity of experimental data. Approaches like multitask modeling and collaborative filtering seek to improve model accuracy by leveraging results from multiple targets, but are limited because different compounds are measured with different assays, leading to sparse data matrices. Profile-QSAR (pQSAR) 2.0 addresses this problem by fitting a series of partial least squares models for each target, using as features the predictions from single-task models on the remaining targets. Here, this method has been shown to produce better results than single task and multitask models. However, the factors determining the success of pQSAR 2.0 have as yet not been characterized. In this paper we examine the experimental conditions that lead to better pQSAR models. We limit the amount of data available to the method by retraining with decreasing amounts of data and explore the model’s ability to generalize to compounds that have never been assayed. Finally, we look at the properties of training data needed to demonstrate pQSAR improvement.
Biological and medical sciences, Computer science↗
Gut microbiome associations with outcome following co-infection with porcine reproductive and respiratory syndrome virus (PRRSV) and porcine circovirus type 2 (PCV2) in pigs immunized with a PRRS modified live virus vaccine
Porcine reproductive and respiratory syndrome virus (PRRSV) and porcine circovirus type 2 (PCV2) are two of the most significant pathogens affecting swine. Co-infections are common and result in respiratory disease and reduced weight gain in growing pigs. Although PRRS modified live virus (MLV) vaccines are widely used to decrease PRRS-associated losses, they are generally considered inadequate for disease control. The gut microbiome provides an alternative strategy to enhance vaccine efficacy and improve PRRS control. The objective of this work was to identify gut microbiome characteristics associated with improved outcome in pigs immunized with a PRRS MLV and co-challenged with PRRSV and PCV2b. Twenty-eight days after vaccination and prior to co-challenge, fecal samples were collected from an experimental population of 50 nursery pigs. At 42 days post-challenge, 20 pigs were retrospectively identified as having high or low growth outcomes during the post-challenge period. Gut microbiomes of the two outcome groups were compared using the Lawrence Livermore Microbial Detection Array (LLMDA) and 16S rDNA sequencing. High growth outcomes were associated with several gut microbiome characteristics, such as increased bacterial diversity, increased Bacteroides pectinophilus, decreased Mycoplasmataceae species diversity, higher Firmicutes:Bacteroidetes ratios, increased relative abundance of the phylum Spirochaetes, reduced relative abundance of the family Lachnospiraceae, and increased Lachnospiraceae species C6A11 and P6B14. Overall, this study identifies gut microbiomes associated with improved outcomes in PRRS vaccinated pigs following a polymicrobial respiratory challenge and provides evidence towards the gut microbiome playing a role in PRRS vaccine efficacy.
59 BASIC BIOLOGICAL SCIENCES↗
Machine Learning Models to Predict Inhibition of the Bile Salt Export Pump
Cholestatic liver injury is frequently associated with drug inhibition of bile salt transporters, such as the bile salt export pump (BSEP). Reliable in silico models to predict BSEP inhibition directly from chemical structures would significantly reduce costs during drug discovery and could help avoid injury to patients. We report our development of classification and regression models for BSEP inhibition with substantially improved performance over previously published models. We assessed the performance effects of different methods of chemical featurization, data set partitioning, and class labeling and identified the methods producing models that generalized best to novel chemical entities.
59 BASIC BIOLOGICAL SCIENCES↗