Foundation model for efficient biological discovery in single-molecule time traces.

The process of removing intronic sequences from a precursor to messenger RNA (pre-mRNA) to yield a mature mRNA transcript via splicing is an integral step in eukaryotic gene expression. Splicing is carried out by a cellular nanomachine called the spliceosome that is composed of RNA components and dozens of proteins. Despite decades of study, many fundamentals of spliceosome function have remained elusive. Recent developments in single-molecule fluorescence microscopy have afforded new tools to better probe the spliceosome and the complex, dynamic process of splicing by direct observation of single molecules. These cutting-edge technologies enable investigators to monitor the dynamics of specific splicing components, whole spliceosomes, and even cotranscriptional splicing within living cells. WIREs RNA 2016, 7:683-701. doi: 10.1002/wrna.1358 For further resources related to this article, please visit the WIREs website.

Techniques such as fluorescence microscopy reveal in three dimensions the complex mechanical structure of the cell cytoskeleton, made up of actin filaments, microtubules and intermediate filaments. Methods to quantify the degree of order in this structure could prove useful in classifying cells and may be related to the structural integrity of the cell. Approximate Entropy (ApEn) is a regularity statistic that has previously been used to quantify randomness in 1D time series data. We have extended the ApEn concept to define a multi-dimensional parameter suitable for application to three dimensional image data. This 3D ApEn parameter was applied to fluorescence microscopy images showing actin filaments in normal and cancerous human prostate cells. Differences between the ApEn calculated from the different images were observed, with cancer cells tending to have higher values indicating a lower degree of order.

Colocalization of differently labeled biomolecules is a valuable tool in fluorescence microscopy and can provide information on biomolecular interactions. With the advent of super-resolution microscopy, colocalization analysis is getting closer to molecular resolution, bridging the gap to other technologies such as fluorescence resonance energy transfer. Among these novel microscopic techniques, single-molecule localization-based super-resolution methods offer the advantage of providing single-molecule coordinates that, rather than intensity information, can be used for colocalization analysis. This requires adapting the existing mathematical algorithms for localization microscopy data. Here, we introduce an algorithm for coordinate-based colocalization analysis which is suited for single-molecule super-resolution data. In addition, we present an experimental configuration for simultaneous dual-color imaging together with a robust approach to correct for optical aberrations with an accuracy of a few nanometers. We demonstrate the potential of our approach for cellular structures and for two proteins binding actin filaments.

Ligand binding of membrane proteins triggers many important cellular signaling events by the lateral aggregation of ligand-bound and other membrane proteins in the plane of the plasma membrane. This local clustering can lead to the co-enrichment of molecules that create an intracellular signal or bring sufficient amounts of activity together to shift an existing equilibrium towards the execution of a signaling event. In this way, clustering can serve as a cellular switch. The underlying uneven distribution and local enrichment of the signaling cluster’s constituting membrane proteins can be used as a functional readout. This information is obtained by combining single-molecule fluorescence microscopy with cluster algorithms that can reliably and reproducibly distinguish clusters from fluctuations in the background noise to generate quantitative data on this complex process. Cluster analysis of single-molecule fluorescence microscopy data has emerged as a proliferative field, and several algorithms and software solutions have been put forward. However, in most cases, such cluster algorithms require multiple analysis parameters to be defined by the user, which may lead to biased results. Furthermore, most cluster algorithms neglect the individual localization precision connected to every localized molecule, leading to imprecise results. Bayesian cluster analysis has been put forward to overcome these problems, but so far, it has entailed high computational cost, increasing runtime drastically. Finally, most software is challenging to use as they require advanced technical knowledge to operate. Here we combined three advanced cluster algorithms with the Bayesian approach and parallelization in a user-friendly GUI and achieved up to an order of magnitude faster processing than for previous approaches. Our work will simplify access to a well-controlled analysis of clustering data generated by SMLM and significantly accelerate data processing. The inclusion of a simulation mode aids in the design of well-controlled experimental assays.

Finding joinable tables in data lakes is key procedure in many applications such as data integration, data augmentation, data analysis, and data market. Traditional approaches that find equi-joinable tables are unable to deal with misspellings and different formats, nor do they capture any semantic joins. In this paper, we propose PEXESO, a framework for joinable table discovery in data lakes. We target the case when textual values are embedded as high-dimensional vectors and columns are joined upon similarity predicates on high-dimensional vectors, hence to address the limitations of equi-join approaches and identify more meaningful results. To efficiently find joinable tables with similarity, we propose a block-and-verify method that utilizes pivot-based filtering. A partitioning technique is developed to cope with the case when the data lake is large and cannot fit in main memory. An experimental evaluation on real datasets shows that our solution identifies substantially more tables than equi-joins and outperforms other similarity-based options, and the join results are useful in data enrichment for machine learning tasks. The experiments also demonstrate the efficiency of the proposed method.

Amyotrophic lateral sclerosis (ALS) is a severe motor neuron disease, with most sporadic cases lacking clear genetic causes. Abnormal pre-mRNA splicing is a fundamental mechanism in neurodegenerative diseases. For example, TAR DNA-binding protein 43 (TDP-43) loss-of-function (LOF) causes widespread RNA mis-splicing events in ALS. Additionally, splicing mutations are major contributors to neurological disorders. However, the role of intronic variants driving RNA mis-splicing in ALS remains poorly understood. To address this, we developed Spliformer to predict RNA splicing. Spliformer is a transformer-based deep learning model trained and tested on splicing events from the GENCODE database, as well as RNA-seq data from blood and central nervous system tissues. We benchmarked Spliformer against SpliceAI and Pangolin using testing datasets and paired whole-genome sequencing (WGS) with RNA-seq data. We further developed the Spliformer-motif model to identify splicing regulatory motifs. We analyzed Clinvar dataset to identify the link of splicing variants with disease pathogenicity. Additionally, we analyzed WGS data of ALS patients and controls to identify common intronic splicing variants linked to ALS risk or disease phenotypes. We also profiled rare intronic splicing variants in ALS patients to identify known or novel ALS-associated genes. Minigene assays were employed to validate candidate splicing variants. Finally, we measured spine density in neurons with a specific gene knockdown or those expressing a TDP-43 disease-causing mutant. Spliformer accurately predicts the possibilities of a nucleotide within a pre-mRNA sequence being a splice donor, acceptor, or neither. Spliformer outperformed SpliceAI and Pangolin in both speed and accuracy in tested splicing events and/or paired WGS/RNA-seq data. Spliformer-motif successfully identified canonical and novel splicing regulatory motifs. In Clinvar dataset, splicing variants are highly related to disease pathogenicity. Genome-wide analyses of common intronic splicing variants nominated one variant linked to ALS progression. Deep learning analyses of WGS data from 1,370 ALS patients revealed rare splicing variants in reported ALS genes (such as PTPRN2 and CFAP410, validated through minigene assays and RNA-seq), and TDP-43 LOF related RNA mis-splicing genes (such as PTPRD). Further genetic analysis and minigene assays nominated PCP4 and TMEM63A as ALS-associated genes. Functional assays demonstrated that PCP4 is critical for maintaining spine density and can rescue spine loss in neurons expressing a disease-causing TDP-43 mutant. In summary, we developed Spliformer and Spliformer-motif that accurately predict and interpret pre-mRNA splicing. Our findings highlight an intronic genetic mechanism driving RNA mis-splicing in ALS and nominate PCP4 as an ALS-associated gene.

Visualization of high-dimensional data such as text documents is widely applicable. The traditional means is to find an appropriate embedding of the high-dimensional representation in a low-dimensional visualizable space. As topic modeling is a useful form of dimensionality reduction that preserves the semantics in documents, recent approaches aim for a visualization that is consistent with both the original word space, as well as the semantic topic space. In this paper, we address the semantic visualization problem. Given a corpus of documents, the objective is to simultaneously learn the topic distributions as well as the visualization coordinates of documents. We propose to develop a semantic visualization model that approximates L2-normalized data directly. The key is to associate each document with three representations: a coordinate in the visualization space, a multinomial distribution in the topic space, and a directional vector in a high-dimensional unit hypersphere in the word space. We join these representations in a unified generative model, and describe its parameter estimation through variational inference. Comprehensive experiments on real-life text datasets show that the proposed method outperforms the existing baselines on objective evaluation metrics for visualization quality and topic interpretability.

Multi-wavelength single-molecule fluorescence analysis of transcription mechanisms

In this work, we release COVID-Twitter-BERT (CT-BERT), a transformer-based model, pretrained on a large corpus of Twitter messages on the topic of COVID-19. Our model shows a 10-30% marginal improvement compared to its base model, BERT-Large, on five different classification datasets. The largest improvements are on the target domain. Pretrained transformer models, such as CT-BERT, are trained on a specific target domain and can be used for a wide variety of natural language processing tasks, including classification, question-answering and chatbots. CT-BERT is optimised to be used on COVID-19 content, in particular social media posts from Twitter.

Multimodal image fusion aims to integrate information from different imaging techniques to produce a comprehensive, detail-rich single image for downstream vision tasks. Existing methods based on local convolutional neural networks (CNNs) struggle to capture global features efficiently, while Transformer-based models are computationally expensive, although they excel at global modeling. Mamba addresses these limitations by leveraging selective structured state space models (S4) to effectively handle long-range dependencies while maintaining linear complexity. In this paper, we propose FusionMamba, a novel dynamic feature enhancement framework that aims to overcome the challenges faced by CNNs and Vision Transformers (ViTs) in computer vision tasks. The framework improves the visual state-space model Mamba by integrating dynamic convolution and channel attention mechanisms, which not only retains its powerful global feature modeling capability, but also greatly reduces redundancy and enhances the expressiveness of local features. In addition, we have developed a new module called the dynamic feature fusion module (DFFM). It combines the dynamic feature enhancement module (DFEM) for texture enhancement and disparity perception with the cross-modal fusion Mamba module (CMFM), which focuses on enhancing the inter-modal correlation while suppressing redundant information. Experiments show that FusionMamba achieves state-of-the-art performance in a variety of multimodal image fusion tasks as well as downstream experiments, demonstrating its broad applicability and superiority.

High-resolution fluorescence microscopy dataset for neuron segmentation and morphological analysis

Single-Molecule Fluorescence Studies of Membrane Transporters Using Total Internal Reflection Microscopy.

Detecting multi-word expressions (MWEs) is essential in natural language processing (NLP) for machine translation, sentiment analysis, and information retrieval. We evaluated the performance of different deep learning models on MWE detection using samples from the major consumer electronics product retailer corpus. We explored transformer-based models such as bidirectional encoder representations from transformers (BERT), BERT-Conditional Random Field (CRF), and robustly optimized BERT variant (RoBERTa)-CRF, alongside a recurrent neural network-based model such as long short-term memory (LSTM)-CRF, using a sample of 10,000 rows, with the CRF layer incorporated in model variations. The transformer-based model with a CRF layer, RoBERTa-CRF, showed the best performance by leveraging the contextual embeddings and sequential dependency modeling. The study results underscore the importance of integrating contextual and sequential deep learning techniques to improve MWE detection and provide a basis for developing robust and scalable systems for diverse linguistic tasks.

The single-molecule approach seeks to understand molecular mechanisms by observing biomolecular processes at the level of individual molecules. These methods have led to a developing understanding that for many processes, a diversity of behaviours will be observed, representing a multitude of pathways. This realisation necessitates that an adequate number of observations are recorded to fully characterise this diversity. The requirement for large numbers of observations to adequately sample distributions, subpopulations, and rare events presents a significant challenge for single-molecule techniques, which by their nature do not typically provide very high throughput. This review will discuss many developing techniques which address this issue by combining nanolithographic approaches, such as zero-mode waveguides and DNA curtains, with single-molecule fluorescence microscopy, and by drastically increasing throughput of force-based approaches such as magnetic tweezers and laminar-flow techniques. These methods not only allow the collection of large volumes of single-molecule data in single experiments, but have also made improvements to ease-of-use, accessibility, and automation of data analysis.

Whole genome and exome sequencing (WGS/WES) are the most popular next‐generation sequencing (NGS) methodologies and are at present often used to detect rare and common genetic variants of clinical significance. We emphasize that automated sequence data processing, management, and visualization should be an indispensable component of modern WGS and WES data analysis for sequence assembly, variant detection (SNPs, SVs), imputation, and resolution of haplotypes. In this manuscript, we present a newly developed findable, accessible, interoperable, and reusable (FAIR) bioinformatics‐genomics pipeline Java based Whole Genome/Exome Sequence Data Processing Pipeline (JWES) for efficient variant discovery and interpretation, and big data modeling and visualization. JWES is a cross‐platform, user‐friendly, product line application, that entails three modules: (a) data processing, (b) storage, and (c) visualization. The data processing module performs a series of different tasks for variant calling, the data storage module efficiently manages high‐volume gene‐variant data, and the data visualization module supports variant data interpretation with Circos graphs. The performance of JWES was tested and validated in‐house with different experiments, using Microsoft Windows, macOS Big Sur, and UNIX operating systems. JWES is an open‐source and freely available pipeline, allowing scientists to take full advantage of all the computing resources available, without requiring much computer science knowledge. We have successfully applied JWES for processing, management, and gene‐variant discovery, annotation, prediction, and genotyping of WGS and WES data to analyze variable complex disorders. In summary, we report the performance of JWES with some reproducible case studies, using open access and in‐house generated, high‐quality datasets.