Critical transitions and tipping phenomena between two meta-stable states in stochastic dynamical systems are a scientific issue. In this work, we expand the methodology of identifying the most probable transition pathway between two meta-stable states with Onsager-Machlup action functional, to investigate the evolutionary transition dynamics between two meta-stable invariant sets with Schrödinger bridge. In contrast to existing methodologies such as statistical analysis, bifurcation theory, information theory, statistical physics, topology, and graph theory for early warning indicators, we introduce a novel framework on Early Warning Signals (EWS) within the realm of probability measures that align with the entropy production rate. To validate our framework, we apply it to the Morris-Lecar model and investigate the transition dynamics between a meta-stable state and a stable invariant set (the limit cycle or homoclinic orbit) under various conditions. Additionally, we analyze real Alzheimer’s data from the Alzheimer’s Disease Neuroimaging Initiative database to explore EWS indicating the transition from healthy to pre-AD states. This framework not only expands the transition pathway to encompass measures between two specified densities on invariant sets, but also demonstrates the potential of our early warning indicators for complex diseases.
It is challenging to identify comorbidity patterns and mechanistically investigate disease associations based on health-related data that are often sparse, large-scale, and multimodal. Adopting a systems biology approach, embedding-based algorithms provide a new perspective to examine diseases under a unified framework by mapping diseases into a high-dimensional space as embedding vectors. These vectors and their constituted disease space encode pathological information and enable a quantitative and systemic measurement of the similarity between any pair of diseases, opening up an avenue for numerous types of downstream analyses. Here, we exemplify its potential through applications in discovering hidden disease associations, assisting in genetic parameter estimation, facilitating data-driven disease classifications, and transforming genetic association studies of diseases in consideration of comorbidities. While underscoring the power and versatility of this approach, we also discuss the challenges posed by medical context, requirements of online training and result validation, and research opportunities in constructing foundation models from multimodal disease data. With continued innovation and exploration, disease embedding has the potential to transform the fields of disease association analysis and even pathology studies by providing a holistic representation of patient health status.
Terahertz (THz) electromagnetic fields are increasingly recognized for their crucial roles in various aspects of medical research and treatment. Recent computational studies have demonstrated that THz waves can modulate ion channel function by interacting with either the channel proteins or the bound ions through distinct mechanisms. Here, we outline a universal simulation protocol to identify the THz frequencies that may affect ion channels, which consists of frequency spectrum analysis and ion conductance analysis. Following this protocol, we studied the effect of the THz field on a CaV channel and found a broad frequency band in the 1-20 THz range. We believe that this protocol, along with the identified characteristic frequencies, will provide a theoretical foundation for future terahertz experimental studies.
Intrinsically disordered proteins (IDP) are highly dynamic, and the effective characterization of IDP conformations is still a challenge. Here, we analyze the chain topology of IDPs and focus on the physical link of the IDP chain, that is, the entanglement between two segments along the IDP chain. The Gauss linking number of two segments throughout the IDP chain is systematically calculated to analyze the physical link. The crossing points of physical links are identified and denoted as link nodes. We notice that the residues involved in link nodes tend to have lower root mean square fluctuation (RMSF), that is, the entanglement of the IDP chain may affect its conformation fluctuation. Moreover, the evolution of the physical link is considerably slow with a timescale of hundreds of nanoseconds. The essential conformation evolution may be depicted on the basis of chain topology.
Hypermutable cancers create opportunities for the development of various immunotherapies, such as immune checkpoint blockade (ICB) therapy. However, emergent studies have revealed that many hypermutated tumors have poor prognosis due to heterogeneous tumor antigen landscapes, yet the underlying mechanisms remain poorly understood. To understand the mechanisms that govern the responses to therapies, we develop mathematical models to explore the impact of combining chemotherapy and ICB therapy on heterogeneous tumors. Our results uncover how chemotherapy reduces antigenic heterogeneity, creating improved immunological conditions within tumors, which, in turn, enhances the therapeutic effect when combined with ICB. Furthermore, our results show that the recovery of the immune system after chemotherapy is crucial for enhancing the response to chemo-ICB combination therapy.
Next-generation sequencing data are widely utilised for various downstream applications in bioinformatics and numerous techniques have been developed for PCR-deduplication and error-correction to eliminate bias and errors introduced during the sequencing. This study first-time provides a joint overview of recent advances in PCR-deduplication and error-correction on short reads. In particular, we utilise UMI-based PCR-deduplication strategies and sequencing data to assess the performance of the solely-computational PCR-deduplication approaches and investigate how error correction affects the performance of PCR-deduplication. Our survey and comparative analysis reveal that the deduplicated reads generated by the solely-computational PCR-deduplication and error-correction methods exhibit substantial differences and divergence from the sets of reads obtained by the UMI-based deduplication methods. The existing solely-computational PCR-deduplication and error-correction tools can eliminate some errors but still leave hundreds of thousands of erroneous reads uncorrected. All the error-correction approaches raise thousands or more new sequences after correction which do not have any benefit to the PCR-deduplication process. Based on our findings, we discuss future research directions and make suggestions for improving existing computational approaches to enhance the quality of short-read sequencing data.
Long noncoding RNAs (lncRNAs) are crucial in gene regulation, chromatin architecture, and cellular differentiation, playing significant roles in various diseases and serving as potential biomarkers and therapeutic targets. Understanding their precise subcellular localization is essential for elucidating their functions in biological pathways. Current methods for predicting lncRNA subcellular localization face challenges in capturing long-range interactions within sequences. Deep learning models often struggle with feature extraction that adequately represents these distant dependencies, leading to limited predictive accuracy. We develop Loc4Lnc, a deep learning framework for predicting lncRNA subcellular localization. The model integrates convolutional layers and transformer blocks to effectively capture both local sequence motifs and long-range dependencies within RNA sequences, followed by classification using TextCNN. Using the RNALocate v2.0 database, we constructed a benchmark dataset covering five subcellular locations (cytoplasm, nucleus, cytosol, chromatin, and exosome). The performance of the model is evaluated against existing feature extraction methods and existing predictors. Results of the Loc4Lnc study demonstrate significant improvements in predicting lncRNA subcellular localization. The model achieved a prediction accuracy of 0.636 on an independent test set, outperforming existing methodologies. Comparative evaluations showed that it consistently surpassed traditional feature extraction methods and state-of-the-art predictors, highlighting its robustness and effectiveness in accurately classifying lncRNAs across five distinct subcellular locations. Loc4Lnc effectively captures long-range interactions and optimizes information flow between distal elements, providing an effective predictive tool for the subcellular localization of lncRNAs and laying the foundation for future research on the regulation of gene expression and cellular functions by lncRNAs.
Recent advances in single-cell DNA methylation have provided unprecedented opportunities to explore cellular epigenetic differences with maximal resolution. A common workflow for single-cell DNA methylation analysis is binning the genome into multiple regions and computing the average methylation level within each region. In this process, imputing not available (NA) values which are caused by the limited number of captured methylation sites is a necessary preprocessing step for downstream analyses. Existing studies have employed several simple imputation methods (such as zeros imputation or means imputation), however, there is a lack of theoretical studies or benchmark tests of these approaches. Through both experiments and theoretical analysis, we found that using the medians to impute NA values can effectively and simply reflect the methylation state of the NA values, providing an accurate foundation for downstream analyses.
Alzheimer's disease (AD) is a common neurodegenerative disorder nowadays. Amyloid-beta (Aβ) and tau proteins are among the main contributors to the AD progression. In AD, Aβ proteins clump together to form plaques and disrupt cell functions. On the other hand, the abnormal chemical change in the brain helps to build sticky tau tangles that block the neuron's transport system. Astrocytes generally maintain a healthy balance in the brain by clearing the Aβ plaques (toxic Aβ). However, overactivated astrocytes release chemokines and cytokines in the presence of Aβ and react to pro-inflammatory cytokines, further increasing the production of Aβ. In this study, we construct a mathematical model that can capture astrocytes' dual behavior. Furthermore, we reveal that the disease progression depends on the current time instance and the disease's earlier status, called the “memory effect,” making non-Markovian processes an appropriate approach. We consider a fractional order network mathematical model to capture the influence of such memory effects on AD progression. We have integrated brain connectome data into the model and studied the memory effect, the dual role of astrocytes, and the brain's neuronal damage. Based on the pathology, primary, secondary, and mixed tauopathies parameters are considered in the model. Due to the mixed tauopathy, different brain nodes or regions in the brain connectome accumulate different toxic concentrations of Aβ and tau proteins. Finally, we explain how the memory effect can slow down the propagation of such toxic proteins in the brain, decreasing the rate of neuronal damage.
Our understanding of intricate biological systems has been completely transformed by the development of multi-omics approaches, which entail the simultaneous study of several different molecular data types. However, there are many obstacles to overcome when analyzing multi-omics data, including the requirement for sophisticated data processing and analysis tools. The integration of multi-omics research with artificial intelligence (AI) has the potential to fundamentally alter our understanding of biological systems. AI has emerged as an effective tool for evaluating complicated data sets. The application of AI and data processing techniques in multi-omics analysis is explored in this study. The present study articulates the diverse categories of information generated by multi-omics methodologies and the intricacies involved in managing and merging these datasets. Additionally, it looks at the various AI techniques—such as machine learning, deep learning, and neural networks—that have been created for multi-omics analysis. The assessment comes to the conclusion that multi-omics analysis has a lot of potential to change with the integration of AI and data processing techniques. AI can speed up the discovery of new biomarkers and therapeutic targets as well as the advancement of personalized medicine strategies by enabling the integration and analysis of massive and complicated data sets. The necessity for high-quality data sets and the creation of useful algorithms and models are some of the difficulties that come with using AI in multi-omics study. In order to fully exploit the promise of AI in multi-omics analysis, more study in this area is required.