In Silico Study of the Acquired Resistance Caused by the Secondary Mutations of KRAS G12C Protein Using Long Time Molecular Dynamics Simulation and Markov State Model Analysis.

Kirsten rat sarcoma viral oncogene homolog (KRAS) is a small GTPase protein which plays an important role in the treatment of KRAS mutant cancers. The FDA-approved AMG510 and MRTX849 (phase III clinical trials) are two potent KRASG12C-selective inhibitors that target KRAS G12C. However, the drug resistance caused by the second-site mutation in KRAS has emerged, and the mechanisms of drug resistance at atom level are still unclear. To clarify the mechanisms of drug resistance, we conducted long time molecular dynamics simulations (75 µs in total) to study the structural and energetic features of KRAS G12C and its four drug resistant variants to inhibitors. The combined binding free energy calculation and protein-ligand interaction fingerprint revealed that these second-site mutations indeed caused KRAS to produce different degrees of resistance to AMG510 and MRTX849. Furthermore, Markov State Models and 2D-free energy landscapes analysis revealed the difference in conformational changes of mutated KRAS bound with and without inhibitors. Furthermore, the comparative analysis of these systems showed that there were differences in their allosteric signal pathways. These findings provide the molecular mechanism of drug resistance, which helps to guide novel KRAS G12C inhibitor design to overcome drug resistance.

Cost-Effectiveness of Mobile Health-Based Integrated Care for Atrial Fibrillation: Model Development and Data Analysis.

BACKGROUND: Mobile health (mHealth) technology is increasingly used in disease management. Using mHealth tools to integrate and streamline care has improved clinical outcomes of patients with atrial fibrillation (AF). OBJECTIVE: The aim of this study was to investigate the potential clinical and health economic outcomes of mHealth-based integrated care for AF from the perspective of a public health care provider in China. METHODS: A Markov model was designed to compare outcomes of mHealth-based care and usual care in a hypothetical cohort of patients with AF in China. The time horizon was 30 years with monthly cycles. Model outcomes measured were direct medical cost, quality-adjusted life years (QALYs), and incremental cost-effectiveness ratio (ICER). Sensitivity analyses were performed to examine the robustness of the base-case results. RESULTS: In the base-case analysis, mHealth-based care gained higher QALYs of 0.0730 with an incurred cost of US $1090. Using US $33,438 per QALY (three times the gross domestic product) as the willingness-to-pay threshold, mHealth-based care was cost-effective, with an ICER of US $14,936 per QALY. In one-way sensitivity analysis, no influential factor with a threshold value was identified. In probabilistic sensitivity analysis, mHealth-based care was accepted as cost-effective in 92.33% of 10,000 iterations. CONCLUSIONS: This study assessed the expected cost-effectiveness of applying mHealth-based integrated care for AF according to a model-based health economic evaluation. The exploration suggested the potential cost-effective use of mHealth apps in streamlining and integrating care via the Atrial fibrillation Better Care (ABC) pathway for AF in China. Future economic evaluation alongside randomized clinical trials is highly warranted to verify the suggestion and investigate affecting factors such as geographical variations in patient characteristics, identification of subgroups, and constraints on local implementation.

Do we need different machine learning algorithms for QSAR modeling? A comprehensive assessment of 16 machine learning algorithms on 14 QSAR data sets.

Although a wide variety of machine learning (ML) algorithms have been utilized to learn quantitative structure-activity relationships (QSARs), there is no agreed single best algorithm for QSAR learning. Therefore, a comprehensive understanding of the performance characteristics of popular ML algorithms used in QSAR learning is highly desirable. In this study, five linear algorithms [linear function Gaussian process regression (linear-GPR), linear function support vector machine (linear-SVM), partial least squares regression (PLSR), multiple linear regression (MLR) and principal component regression (PCR)], three analogizers [radial basis function support vector machine (rbf-SVM), K-nearest neighbor (KNN) and radial basis function Gaussian process regression (rbf-GPR)], six symbolists [extreme gradient boosting (XGBoost), Cubist, random forest (RF), multiple adaptive regression splines (MARS), gradient boosting machine (GBM), and classification and regression tree (CART)] and two connectionists [principal component analysis artificial neural network (pca-ANN) and deep neural network (DNN)] were employed to learn the regression-based QSAR models for 14 public data sets comprising nine physicochemical properties and five toxicity endpoints. The results show that rbf-SVM, rbf-GPR, XGBoost and DNN generally illustrate better performances than the other algorithms. The overall performances of different algorithms can be ranked from the best to the worst as follows: rbf-SVM > XGBoost > rbf-GPR > Cubist > GBM > DNN > RF > pca-ANN > MARS > linear-GPR ≈ KNN > linear-SVM ≈ PLSR > CART ≈ PCR ≈ MLR. In terms of prediction accuracy and computational efficiency, SVM and XGBoost are recommended to the regression learning for small data sets, and XGBoost is an excellent choice for large data sets. We then investigated the performances of the ensemble models by integrating the predictions of multiple ML algorithms. The results illustrate that the ensembles of two or three algorithms in different categories can indeed improve the predictions of the best individual ML algorithms.

Artificial intelligence-based collaborative filtering method with ensemble learning for personalized lung cancer medicine without genetic sequencing.

In personalized medicine, many factors influence the choice of compounds. Hence, the selection of suitable medicine for patients with non-small-cell lung cancer (NSCLC) is expensive. To shorten the decision-making process for compounds, we propose a computationally efficient and cost-effective collaborative filtering method with ensemble learning. The ensemble learning is used to handle small-sample sizes in drug response datasets as the typical number of patients in a cancer dataset is very small. Moreover, the proposed method can be used to identify the most suitable compounds for patients without genetic data. To the best of our knowledge, this is the first method to provide effective recommendations without genetic data. We also constructed a reliable dataset that includes eight NSCLC cell lines and ten compounds that have been approved by the Food and Drug Administration. With the new dataset, the experimental results demonstrated that the dataset shift phenomenon that commonly occurs in practical biomedical data does not occur in this problem. The experimental results demonstrated that our proposed method can outperform two state-of-the-art recommender system techniques on both the NCI60 dataset and our new dataset. Our model can be applied to the prediction of drug sensitivity with less labor-intensive experiments in the future.

Neurostructural correlates of optimism: Gray matter density in the putamen predicts dispositional optimism in late adolescence.

Dispositional optimism reflects one's generalized positive expectancies for future outcomes and plays a crucial role in personal developmental outcomes and health (e.g., counteracting related mental disorders such as depression and anxiety). Increasing evidence has suggested that extraversion is an important personality factor contributing to dispositional optimism. However, less is known about the association between dispositional optimism and brain structure and the role of extraversion in this association. Here, we examined these issues in 231 healthy high school students aged 16 to 20 years (110 males, mean age = 18.48 years, SD = 0.54) by estimating regional gray matter density (rGMD) using a voxel-based morphometry method via structural magnetic resonance imaging. Whole-brain regression analyses revealed a significant positive correlation between dispositional optimism and the rGMD of the bilateral putamen after adjusting for age, sex, family socioeconomic status (SES), general intelligence, and total gray matter volume (TGMV). Moreover, prediction analyses using fourfold balanced cross-validation combined with linear regression confirmed a significant connection between dispositional optimism and putamen density after adjusting for age, sex, and family SES. More importantly, subsequent mediation analysis showed that extraversion may account for the association between putamen density and dispositional optimism after adjusting for age, sex, family SES, general intelligence, TGMV, and the other four Big Five personality traits. Taken together, the current study provides new evidence regarding the neurostructural basis underlying dispositional optimism in adolescents and underscores the importance of extraversion as an essential personality factor for dispositional optimism acquisition.

The role of bone mineral density in therapeutic decision-making using the Fracture Risk Assessment Tool (FRAX): a sub-study of the Taiwan OsteoPorosis Survey (TOPS).

Fracture Risk Assessment Tool (FRAX)-based intervention threshold (IT) is widely applied for treatment decision-making; however, an IT based on FRAX without the measurement of bone mass density (BMD) has not been validated. The study demonstrated that estimates of fracture risk by FRAX without BMD were higher than those by FRAX with BMD in women with old age. INTRODUCTION: BMD is an integral component for bone strength assessment, but age-specific impacts of BMD on fracture risk assessment and therapeutic decision-making remained unclear. We aimed to investigate whether using BMD measurement changed the interpretation of the FRAX-based fracture probability assessment and treatment decision. METHODOLOGY: The database was provided by the Taiwanese Osteoporosis Association (TOA) which conducted a nationwide survey of BMD. We calculated the 10-year major and hip fracture probabilities using the FRAX for each participant, either with (FRAX + BMD) or without BMD (FRAX - BMD). Age-specific individual intervention thresholds (IITs) were established using the FRAX-based fracture risk, equivalent to a woman with a prior fracture. Participants whose FRAX scores of major fracture were greater than or equal to their IITs were deemed suitable for therapeutic intervention. RESULTS: A total of 14,007 postmenopausal women were enrolled. Compared with FRAX + BMD, FRAX - BMD predicted lower FRAX scores in major and hip fractures in subjects aged 40-60 years; however, FRAX - BMD estimated higher risks for those aged 61-90 years. The therapeutic decision using FRAX - BMD was concordant to that using FRAX + BMD in 90.5% of the subjects, especially in the younger age group (40-70 years). FRAX - BMD identified more treatment candidates (7.7-16.4%) among those aged 71-90 years. CONCLUSIONS: The FRAX scores are influenced by age, irrespective of the consideration of BMD. FRAX - BMD is able to identify more subjects for therapeutic intervention in the elderly population. We should reconsider the role of BMD at different ages for therapeutic decision-making.