Abstract
Study of human disease remains challenging due to convoluted disease etiologies and complex molecular mechanisms at genetic, genomic, and proteomic levels. Many machine learning-based ...methods have been developed and widely used to alleviate some analytic challenges in complex human disease studies. While enjoying the modeling flexibility and robustness, these model frameworks suffer from non-transparency and difficulty in interpreting each individual feature due to their sophisticated algorithms. However, identifying important biomarkers is a critical pursuit towards assisting researchers to establish novel hypotheses regarding prevention, diagnosis and treatment of complex human diseases. Herein, we propose a Permutation-based Feature Importance Test (PermFIT) for estimating and testing the feature importance, and for assisting interpretation of individual feature in complex frameworks, including deep neural networks, random forests, and support vector machines. PermFIT (available at
https://github.com/SkadiEye/deepTL
) is implemented in a computationally efficient manner, without model refitting. We conduct extensive numerical studies under various scenarios, and show that PermFIT not only yields valid statistical inference, but also improves the prediction accuracy of machine learning models. With the application to the Cancer Genome Atlas kidney tumor data and the HITChip atlas data, PermFIT demonstrates its practical usage in identifying important biomarkers and boosting model prediction performance.
•We devise a reinforcement learning-based framework to relocate the POS more adaptively.•Based on different severity degree of environmental changes, three change response methods are integrated to ...predict the POS.•The proposed algorithm learns the dynamic environment and determines the appropriate prediction-based strategy.
Dynamic Multi-objective Optimization Problem (DMOP) is emerging in recent years as a major real-world optimization problem receiving considerable attention. Tracking the movement of Pareto front efficiently and effectively over time has been a central issue in solving DMOPs. In this paper, a reinforcement learning-based dynamic multi-objective evolutionary algorithm, called RL-DMOEA, which seamlessly integrates reinforcement learning framework and three change response mechanisms, is proposed for solving DMOPs. The proposed algorithm relocates the individuals based on the severity degree of environmental changes, which is estimated through the corresponding changes in the objective space of their decision variables. When identifying different severity degree of environmental changes, the proposed RL-DMOEA approach can learn better evolutionary behaviors from environment information, based on which apply the appropriate response mechanisms. Specifically, these change response mechanisms including the knee-based prediction, center-based prediction and indicator-based local search, are devised to promote both convergence and diversity of the algorithm under different severity of environmental changes. To verify this idea, the proposed RL-DMOEA is evaluated on CEC 2015 test problems involving various problem characteristics. Empirical studies on chosen state-of-the-art designs validate that the proposed RL-DMOEA is effective in addressing the DMOPs.
A misdiagnosis of pulmonary embolism (PE) can have severe consequences such as disability or death. It's crucial to accurately identify key clinical features of PE in clinical practice to promptly ...identify potential PE patients who may present asymptomatically, and to prevent misdiagnosing PE as asthma exacerbation in patients with symptoms like dyspnea or chest pain. However, reliably identifying these important features can be challenging due to many factors influencing the likelihood of PE development in complex fashions (e.g., the interactions among these factors). To address this difficulty, we presented an effective framework using the deep neural network (DNN) model and the permutation-based feature importance test (PermFIT) procedure, i.e., PermFIT-DNN. We applied the PermFIT-DNN framework to the analysis of data from a PE study for asthma exacerbation patients. Our analysis results show that the PermFIT-DNN framework can robustly identify key features for classifying PE status. The important features identified can also aid in accurately predicting the PE risk.
The repair of large bone defects is a great challenge in clinical practice. In this study, copper-loaded-ZIF-8 nanoparticles and poly (lactide-co-glycolide) (PLGA) were combined to fabricate porous ...PLGA/Cu(I)@ZIF-8 scaffolds using three-dimensional printing technology for infected bone repair.
The surface morphology of PLGA/Cu(I)@ZIF-8 scaffolds was investigated by transmission electron microscopy and scanning electron microscopy. The PLGA/Cu(I)@ZIF-8 scaffolds were co-cultured with bacteria to determine their antibacterial properties, and with murine mesenchymal stem cells (MSCs) to explore their biocompatibility and osteoconductive properties. The bioactivity of the PLGA/Cu(I)@ZIF-8 scaffolds was evaluated by incubating in simulated body fluid.
The results revealed that the PLGA/Cu(I)@ZIF-8 scaffolds had porosities of 80.04 ± 5.6% and exhibited good mechanical properties. When incubated with H
O
, Cu(I)@ZIF-8 nanoparticles resulted generated reactive oxygen species, which contributed to their antibacterial properties. The mMSCs cultured on the surface of PLGA/Cu(I)@ZIF-8 scaffolds were well-spread and adherent with a high proliferation rate, and staining with alkaline phosphatase and alizarin red was increased compared with the pure PLGA scaffolds. The mineralization assay showed an apatite-rich layer was formed on the surface of PLGA/Cu(I)@ZIF-8 scaffolds, while there was hardly any apatite on the surface of the PLGA scaffolds. Additionally, in vitro, Staphylococcus aureus cultured on the PLGA/Cu(I)@ZIF-8 scaffolds were almost all dead, while in vivo inflammatory cell infiltration and bacteria numbers were dramatically reduced in infected rats implanted with PLGA/Cu@ZIF-8 scaffolds.
All these findings demonstrate that PLGA/Cu(I)@ZIF-8 scaffolds possess excellent antibacterial and osteoconductive properties, as well as good biocompatibility and high bioactivity. This study suggests that the PLGA/Cu(I)@ZIF-8 scaffolds could be used as a promising biomaterial for bone tissue engineering, especially for infected bone repair.
Based on Stephen Paget's well-established theory, both cell-autonomous and non-cell-autonomous mechanisms are crucial for metastasis. Although the mitochondrial calcium uniporter (MCU) has been ...suggested to be involved in breast cancer (BC) progression via cell-autonomous mechanisms, whether it assists the metastasis of BC cells through non-cell-autonomous mechanisms remains unclear. This study aimed to demonstrate that the MCU regulates BC metastatic colonization via non-cell-autonomous mechanisms. The results suggested that extracellular vesicles (EVs) derived from MCU-downregulated MDA-MB-231 cells suppressed angiogenesis in the metastatic niche in a nude mouse model, thereby hindering the colonization of BC cells. Mechanistically, we revealed that the MCU negatively correlated with miR-4488 in EVs derived from BC cells. Significantly, miR-4488 was determined to suppress angiogenesis of vascular endothelial cells by directly targeting angiogenic CX3CL1. Furthermore, we identified miR-4488 as being significantly downregulated in serum EVs from patients with triple-negative BC. Hence, this study suggests that MCU-dependent negative sorting of miR-4488 to EVs enhances angiogenesis in the metastatic niche and, thus, favors the metastatic colonization of BC cells.
As a promising tool for identifying genetic markers underlying phenotypic differences, genome-wide association study (GWAS) has been extensively investigated in recent years. In GWAS, detecting ...epistasis (or gene–gene interaction) is preferable over single locus study since many diseases are known to be complex traits. A brute force search is infeasible for epistasis detection in the genome-wide scale because of the intensive computational burden. Existing epistasis detection algorithms are designed for dataset consisting of homozygous markers and small sample size. In human study, however, the genotype may be heterozygous, and number of individuals can be up to thousands. Thus, existing methods are not readily applicable to human datasets. In this article, we propose an efficient algorithm, TEAM, which significantly speeds up epistasis detection for human GWAS. Our algorithm is exhaustive, i.e. it does not ignore any epistatic interaction. Utilizing the minimum spanning tree structure, the algorithm incrementally updates the contingency tables for epistatic tests without scanning all individuals. Our algorithm has broader applicability and is more efficient than existing methods for large sample study. It supports any statistical test that is based on contingency tables, and enables both family-wise error rate and false discovery rate controlling. Extensive experiments show that our algorithm only needs to examine a small portion of the individuals to update the contingency tables, and it achieves at least an order of magnitude speed up over the brute force approach. Contact: xiang@cs.unc.edu
An ENsemble Deep Learning Optimal Treatment (EndLot) approach is proposed for personalized medicine problems. The statistical framework of the proposed method is based on the outcome weighted ...learning (OWL) framework which transforms the optimal decision rule problem into a weighted classification problem. We further employ an ensemble of deep neural networks (DNNs) to learn the optimal decision rule. Utilizing the flexibility of DNNs and the stability of bootstrap aggregation, the proposed method achieves a considerable improvement over existing methods. An R package “ITRlearn” is developed to implement the proposed method. Numerical performance is demonstrated via simulation studies and a real data analysis of the Cancer Cell Line Encyclopedia data.
Methylmercury (MeHg), an environmental toxin, may specifically cause neurological disorders. Recent studies have reported that autophagy can be induced by metals and be involved in metal ...cytotoxicity. However, the role of autophagy in MeHg-induced neurotoxicity remains unknown. Here, we demonstrate that MeHg induces mTOR-independent autophagy through JNK/Vps34 complex pathway, which further promotes autophagosome accumulation and neuronal cell death. In addition to cell death, MeHg increased LC3-II expression in a concentration- and time-dependent manner in neuronal cells; furthermore, western blot analysis of LC3-II expression under baf A1-treated condition indicates that MeHg activates autophagy induction. However, we found lysosomal degradative function was impaired by MeHg. Under this condition, MeHg-activated autophagy induction would elicit autophagosome accumulation and cell death. Consistent with this inference, the autophagy inhibitor decreased the MeHg-induced autophagosome accumulation and neuronal cells death, whereas the autophagy inducers further augmented MeHg cytotoxicity. Then, the mechanism of autophagy induction is investigated. We show that MeHg-induced autophagy is mTOR-independent. Vacuolar protein sorting 34 (Vps34) complex is critical for mTOR-independent autophagy. MeHg induced the interaction between Beclin1 and Vps34 to form Vps34 complex. Importantly, knockdown of Vps34 inhibited autophagy induction by MeHg. Furthermore, we found that JNK, but not p38 or ERK, promoted the formation of Vps34 complex and autophagy induction. Finally, inhibition of JNK or downregulation of Vps34 decreased autophagosome accumulation and alleviated MeHg-induced neuronal cell death. The present study implies that inhibiting JNK/Vps34 complex autophagy induction pathway may be a novel therapeutic approach for the treatment of MeHg-induced neurotoxicity.
Hepatocellular cancer (HCC) has been reported to belong to one of the highly vascularized solid tumours accompanied with angiogenesis of human umbilical vein endothelial cells (HUVECs). KDM5A, an ...attractive drug target, plays a critical role in diverse physiological processes. Thus, this study aims to investigate its role in angiogenesis and underlying mechanisms in HCC. ChIP‐qPCR was utilized to validate enrichment of H3K4me3 and KDM5A on the promotor region of miR‐433, while dual luciferase assay was carried out to confirm the targeting relationship between miR‐433 and FXYD3. Scratch assay, transwell assay, Edu assay, pseudo‐tube formation assay and mice with xenografted tumours were conducted to investigate the physiological function of KDM5A‐miR‐433‐FXYD3‐PI3K‐AKT axis in the progression of HCC after loss‐ and gain‐function assays. KDM5A p‐p85 and p‐AKT were highly expressed but miR‐433 was down‐regulated in HCC tissues and cell lines. Depletion of KDM5A led to reduced migrative, invasive and proliferative capacities in HCC cells, including growth and a lowered HUVEC angiogenic capacity in vitro. Furthermore, KDM5A suppressed the expression of miR‐433 by demethylating H3K4me3 on its promoterregion. miR‐433 negatively targeted FXYD3. Depleting miR‐433 or re‐expressing FXYD3 restores the reduced migrative, invasive and proliferative capacities, and lowers the HUVEC angiogenic capacity caused by silencing KDM5A. Therefore, KDM5A silencing significantly suppresses HCC tumorigenesis in vivo, accompanied with down‐regulated miR‐433 and up‐regulated FXYD3‐PI3K‐AKT axis in tumour tissues. Lastly, KDM5A activates the FXYD3‐PI3K‐AKT axis to enhance angiogenesis in HCC by suppressing miR‐433.
Summary
Retinoic acid‐inducible gene I (RIG‐I) ‐like receptors (RLRs) are found conservatively present in teleost fish. All three members, RIG‐I, MDA5 and LGP2, together with the downstream molecules ...such as MITA, TRAF3 and TBK1, have been identified in a range of fish species. However, it is unexpected that RIG‐I has not been reported in fish of Acanthopterygii, and it would be important to clarify the presence and role of the RIG‐I gene in a broad range of taxa in Teleostei. RLRs in fish can be induced in vivo and in vitro by viral pathogens as well as synthetic dsRNA, poly(I:C), leading to the production of type I interferons (IFNs) and the expression of IFN‐stimulated genes (ISGs). Bacterial pathogens, such as Edwardsiella tarda, and their components, such as lipopolysaccharide are also found to induce the expression of RLRs, and whether such induction was mediated through the direct recognition by RLRs or through crosstalk with other pattern recognition receptors recognizing directly bacterial pathogen‐associated molecular patterns awaits to be investigated. On the other hand, RLR‐activated type I IFN production can be negatively regulated in fish by molecules, such as TBK‐1‐like protein and IRF10, which are found to negatively regulate RIG‐I and MAVS‐activated type I IFN production, and to block MITA or bind ISRE motifs, respectively. It is considered that the evolutionary occurrence of RLRs in fish, and their recognized ligands, especially those from their fish pathogens, as well as the mechanisms involved in the RLR signalling pathways, are of significant interest for further investigation.
RIG‐I‐like receptors (RLRs) are found conservatively present in teleost fish. But RIG‐I has not been reported in fish of Acanthopterygii. The current knowledge on RLRs in fish is reviewed. The mechanisms involved in the RLR signalling pathways are of significant interest for further investigations in fish.