Search

Drug response prediction is hampered by uncertainty in the measures of response and selection of doses. In this study, we propose a probabilistic multi-output model to simultaneously predict all dose–responses and uncover their biomarkers. By describing the relationship between genomic features and chemical properties to every response at every dose, our multi-output Gaussian Process (MOGP) models enable assessment of drug efficacy using any dose–response metric. This approach was tested across two drug screening studies and ten cancer types. Kullback-leibler divergence measured the importance of each feature and identified EZH2 gene as a novel biomarker of BRAF inhibitor response. We demonstrate the effectiveness of our MOGP models in accurately predicting dose–responses in different cancer types and when there is a limited number of drug screening experiments for training. Our findings highlight the potential of MOGP models in enhancing drug development pipelines by reducing data requirements and improving precision in dose–response predictions. [ABSTRACT FROM AUTHOR]

Motivation Personalized cancer treatments require accurate drug response predictions. Existing deep learning methods show promise but higher accuracy is needed to serve the purpose of precision medicine. The prediction accuracy can be improved with not only topology but geometrical information of drugs. Results A novel deep learning methodology for drug response prediction is presented, named Hi-GeoMVP. It synthesizes hierarchical drug representation with multi-omics data, leveraging graph neural networks and variational autoencoders for detailed drug and cell line representations. Multi-task learning is employed to make better prediction, while both 2D and 3D molecular representations capture comprehensive drug information. Testing on the GDSC dataset confirms Hi-GeoMVP's enhanced performance, surpassing prior state-of-the-art methods by improving the Pearson correlation coefficient from 0.934 to 0.941 and decreasing the root mean square error from 0.969 to 0.931. In the case of blind test, Hi-GeoMVP demonstrated robustness, outperforming the best previous models with a superior Pearson correlation coefficient in the drug-blind test. These results underscore Hi-GeoMVP's capabilities in drug response prediction, implying its potential for precision medicine. Availability and implementation The source code is available at https://github.com/matcyr/Hi-GeoMVP [ABSTRACT FROM AUTHOR]

The diagnostic value of audiovisual sexual stimulation (AVSS) for psychogenic erectile dysfunction (ED) is still unclear. We investigated the independent diagnostic value and optimal cut-off parameter of AVSS for psychogenic ED. All participants had received the AVSS test and nocturnal penile tumescence and rigidity (NPTR) monitoring at least twice. ED patients were divided into psychogenic ED and organic ED according to NPTR examination. The diagnostic accuracy of AVSS parameters was evaluated with the receiver operating characteristic (ROC) curve, and the Youden index was employed to determine the optimal diagnostic cut-off values. A total of 346 patients with ED and 60 healthy men were included in this study, among which 162 and 184 cases of psychogenic and organic ED were identified based on NPTR, respectively. When comparing the two ED groups, the area under the curve (AUC) of AVSS parameters was 0.85–0.89. Six-selected AVSS parameters could precisely diagnose psychogenic ED, exhibiting increased diagnostic specificity compared with corresponding sensitivity. When comparing psychogenic ED with the control group, the AUC of the tumescence of the tip was superior to the AUC other parameters (0.81 vs. 0.58, 0.66, 0.59, 0.53, 0.68), and the best determined diagnostic cut-off value was the tumescence of the tip < 29.87%. Independent AVSS could diagnose psychogenic ED objectively and effectively, and its diagnostic value was highest when 1.50% ≤ tumescence of the tip < 29.87%. [ABSTRACT FROM AUTHOR]

Mechanosensitive (MS) ion channels provide efficient molecular mechanism for transducing mechanical forces into intracellular ion fluxes in all kingdoms of life. The mechanosensitive channel of small conductance (MscS) was one of the best-studied MS channels and its homologs (MSL, MscS-like) were widely distributed in cell-walled organisms. However, the origin, evolution and expansion of MSL proteins in plants are still not clear. Here, we identified more than 2100 MSL proteins from 176 plants and conducted a broad-scale phylogenetic analysis. The phylogenetic tree showed that plant MSL proteins were divided into three groups (I, II and III) prior to the emergence of chlorophytae algae, consistent with their specific subcellular localization. MSL proteins were distributed unevenly into each of plant species, and four parallel expansion was identified in angiosperms. In Brassicaceae, most MSL duplicates were derived by whole-genome duplication (WGD)/segmental duplications. Finally, a hypothetical evolutionary model of MSL proteins in plants was proposed based on phylogeny. Our studies illustrate the evolutionary history of the MSL proteins and provide a guide for future functional diversity analyses of these proteins in plants. [ABSTRACT FROM AUTHOR]

The drug response prediction problem arises from personalized medicine and drug discovery. Deep neural networks have been applied to the multi-omics data being available for over 1000 cancer cell lines and tissues for better drug response prediction. We summarize and examine state-of-the-art deep learning methods that have been published recently. Although significant progresses have been made in deep learning approach in drug response prediction, deep learning methods show their weakness for predicting the response of a drug that does not appear in the training dataset. In particular, all the five evaluated deep learning methods performed worst than the similarity-regularized matrix factorization (SRMF) method in our drug blind test. We outline the challenges in applying deep learning approach to drug response prediction and suggest unique opportunities for deep learning integrated with established bioinformatics analyses to overcome some of these challenges. [ABSTRACT FROM AUTHOR]

Despite many strong and tough hydrogels have been fabricated according to the energy dissipating mechanism, they usually lack high resilience due to the presence of large hysteresis. Herein, poly (N‐vinylpyrrolidone) grafted cellulose nanocrystal (CNC‐g‐PVP) was used as special multifunctional physical crosslinkers to fabricate tough and highly resilient nanocomposite hydrogels. CNC‐g‐PVP with varying loading was incorporated into chemically crosslinked polyacrylamide (PAM) networks by in‐situ radical polymerization to give PAM/CNC‐g‐PVP nanocomposite hydrogels. Robust cooperative hydrogen bonds existed between the surface‐grafted PVP chains and the PAM matrix, which could rupture to dissipate energy upon deformation and recover instantly on the removal of stress. This unique energy dissipating mechanism led to excellent mechanical performance of the hydrogels. Their tensile elastic modulus, toughness, and compressive strength are 1.4–1.8, 2.1–3.0, and 1.44–2.73 times of pure PAM hydrogel, respectively. Moreover, the hydrogels exhibit low hysteresis, high resilience (ca. 97%) under cyclic tensile loading‐unloading and good recovery of hysteresis (ca. 90%) under cyclic compressive loading‐unloading. [ABSTRACT FROM AUTHOR]