Bioinformatics Challenges Research Articles

Long-read sequencing transcriptome quantification with lr-kallisto.

RNA abundance quantification has become routine and affordable thanks to high-throughput "short-read" technologies that provide accurate molecule counts at the gene level. Similarly accurate and affordable quantification of definitive full-length, transcript isoforms has remained a stubborn challenge, despite its obvious biological significance across a wide range of problems. "Long-read" sequencing platforms now produce data-types that can, in principle, drive routine definitive isoform quantification. However some particulars of contemporary long-read datatypes, together with isoform complexity and genetic variation, present bioinformatic challenges. We show here, using ONT data, that fast and accurate quantification of long-read data is possible and that it is improved by exome capture. To perform quantifications we developed lr-kallisto, which adapts the kallisto bulk and single-cell RNA-seq quantification methods for long-read technologies.

Single-cell DNA methylation analysis tool Amethyst reveals distinct noncanonical methylation patterns in human glial cells.

Single-cell sequencing technologies have revolutionized biomedical research by enabling deconvolution of cell type-specific properties in highly heterogeneous tissue. While robust tools have been developed to handle bioinformatic challenges posed by single-cell RNA and ATAC data, options for emergent modalities such as methylation are much more limited, impeding the utility of results. Here we present Amethyst, a comprehensive R package for atlas-scale single-cell methylation sequencing data analysis. Amethyst begins with base-level methylation calls and expedites batch integration, doublet detection, dimensionality reduction, clustering, cell type annotation, differentially methylated region calling, and interpretation of results, facilitating rapid data interaction in a local environment. We introduce the workflow using published single-cell methylation human peripheral blood mononuclear cell (PBMC) and human cortex data. We further leverage Amethyst on an atlas-scale brain dataset to describe a noncanonical methylation pattern in human astrocytes and oligodendrocytes, challenging the notion that this form of methylation is principally relevant to neurons in the brain. Tools such as Amethyst will increase accessibility to single-cell methylation data analysis, catalyzing research progress across diverse contexts.

Training Data Diversity Enhances the Basecalling of Novel RNA Modification-Induced Nanopore Sequencing Readouts.

Accurately basecalling sequence backbones in the presence of nucleotide modifications remains a substantial challenge in nanopore sequencing bioinformatics. It has been extensively demonstrated that state-of-the-art basecallers are less compatible with modification-induced sequencing signals. A precise basecalling, on the other hand, serves as the prerequisite for virtually all the downstream analyses. Here, we report that basecallers exposed to diverse training modifications gain the generalizability to analyze novel modifications. With synthesized oligos as the model system, we precisely basecall various out-of-sample RNA modifications. From the representation learning perspective, we attribute this generalizability to basecaller representation space expanded by diverse training modifications. Taken together, we conclude increasing the training data diversity as a novel paradigm for building modification-tolerant nanopore sequencing basecallers.

Improving protein function prediction by learning and integrating representations of protein sequences and function labels.

As fewer than 1% of proteins have protein function information determined experimentally, computationally predicting the function of proteins is critical for obtaining functional information for most proteins and has been a major challenge in protein bioinformatics. Despite the significant progress made in protein function prediction by the community in the last decade, the general accuracy of protein function prediction is still not high, particularly for rare function terms associated with few proteins in the protein function annotation database such as the UniProt. We introduce TransFew, a new transformer model, to learn the representations of both protein sequences and function labels [Gene Ontology (GO) terms] to predict the function of proteins. TransFew leverages a large pre-trained protein language model (ESM2-t48) to learn function-relevant representations of proteins from raw protein sequences and uses a biological natural language model (BioBert) and a graph convolutional neural network-based autoencoder to generate semantic representations of GO terms from their textual definition and hierarchical relationships, which are combined together to predict protein function via the cross-attention. Integrating the protein sequence and label representations not only enhances overall function prediction accuracy, but delivers a robust performance of predicting rare function terms with limited annotations by facilitating annotation transfer between GO terms. https://github.com/BioinfoMachineLearning/TransFew.

Deep learning methods for protein function prediction.

Predicting protein function from protein sequence, structure, interaction, and other relevant information is important for generating hypotheses for biological experiments and studying biological systems, and therefore has been a major challenge in protein bioinformatics. Numerous computational methods had been developed to advance protein function prediction gradually in the last two decades. Particularly, in the recent years, leveraging the revolutionary advances in artificial intelligence (AI), more and more deep learning methods have been developed to improve protein function prediction at a faster pace. Here, we provide an in-depth review of the recent developments of deep learning methods for protein function prediction. We summarize the significant advances in the field, identify several remaining major challenges to be tackled, and suggest some potential directions to explore. The data sources and evaluation metrics widely used in protein function prediction are also discussed to assist the machine learning, AI, and bioinformatics communities to develop more cutting-edge methods to advance protein function prediction.

Integrating Large Language Models in Bioinformatics Education for Medical Students: Opportunities and Challenges.

Large language models (LLMs) offer transformative opportunities in bioinformatics education for medical students by creating interactive experiences. The integration of LLMs enhances educational outcomes through providing accessible code templates, clarifying the function of coding elements, and assisting in error troubleshooting. Here, we demonstrate the practical applications of LLMs with a case study on transcriptome sequencing data processing, a vital component of medical research. However, the reliability of the content that LLMs generate requires rigorous validation. Ensuring the accuracy and appropriateness of the LLM-generated information requires integrating innovative LLMs with traditional educational methods to prepare medical students effectively for future challenges in bioinformatics.

Toward structure–multiple activity relationships (SMARts) using computational approaches: A polypharmacological perspective

In the current era of biological big data, which are rapidly populating the biological chemical space, in silico polypharmacology drug design approaches help to decode structure-multiple activity relationships (SMARts). Current computational methods can predict or categorize multiple properties simultaneously, which aids the generation, identification, curation, prioritization, optimization, and repurposing of molecules. Computational methods have generated opportunities and challenges in medicinal chemistry, pharmacology, food chemistry, toxicology, bioinformatics, and chemoinformatics. It is anticipated that computer-guided SMARts could contribute to the full automatization of drug design and drug repurposing campaigns, facilitating the prediction of new biological targets, side and off-target effects, and drug-drug interactions.

Spectral entropy as a measure of the metaproteome complexity.

The diversity and complexity of the microbiome's genomic landscape are not always mirrored in its proteomic profile. Despite the anticipated proteomic diversity, observed complexities of microbiome samples are often lower than expected. Two main factors contribute to this discrepancy: limitations in mass spectrometry's detection sensitivity and bioinformatics challenges in metaproteomics identification. This study introduces a novel approach to evaluating sample complexity directly at the full mass spectrum (MS1) level rather than relying on peptide identifications. When analyzing under identical mass spectrometry conditions, microbiome samples displayed significantly higher complexity, as evidenced by the spectral entropy and peptide candidate entropy, compared to single-species samples. The research provides solid evidence for the complexity of microbiome in proteomics indicating the optimization potential of the bioinformatics workflow.

A global initiative on addressing bioinformatics' grand challenges.

The Bioinformatics Grand Challenges Consortium (BGCC) is a collaborative effort to address the most pressing challenges in bioinformatics. Initially focusing on education and training, the consortium successfully defined seven key grand challenges and is actively developing actionable solutions for these challenges. Building on this foundation, the BGCC plans to broaden its focus to include additional grand challenges in emerging areas.

Exploiting protein language models for the precise classification of ion channels and ion transporters.

This study introduces TooT-PLM-ionCT, a comprehensive framework that consolidates three distinct systems, each meticulously tailored for one of the following tasks: distinguishing ion channels (ICs) from membrane proteins (MPs), segregating ion transporters (ITs) from MPs, and differentiating ICs from ITs. Drawing upon the strengths of six Protein Language Models (PLMs)-ProtBERT, ProtBERT-BFD, ESM-1b, ESM-2 (650M parameters), and ESM-2 (15B parameters), TooT-PLM-ionCT employs a combination of traditional classifiers and deep learning models for nuanced protein classification. Originally validated on an existing dataset by previous researchers, our systems demonstrated superior performance in identifying ITs from MPs and distinguishing ICs from ITs, with the IC-MP discrimination achieving state-of-the-art results. In light of recommendations for additional validation, we introduced a new dataset, significantly enhancing the robustness and generalization of our models across bioinformatics challenges. This new evaluation underscored the effectiveness of TooT-PLM-ionCT in adapting to novel data while maintaining high classification accuracy. Furthermore, this study explores critical factors affecting classification accuracy, such as dataset balancing, the impact of using frozen versus fine-tuned PLM representations, and the variance between half and full precision in floating-point computations. To facilitate broader application and accessibility, a web server (https://tootsuite.encs.concordia.ca/service/TooT-PLM-ionCT) has been developed, allowing users to evaluate unknown protein sequences through our specialized systems for IC-MP, IT-MP, and IC-IT classification tasks.

U-Net for genomic sequencing: A novel approach to DNA sequence classification

The precise classification of DNA sequences is pivotal in genomics, holding significant implications for personalized medicine. The stakes are particularly high when classifying key genetic markers such as BRAC, related to breast cancer susceptibility; BRAF, associated with various malignancies; and KRAS, a recognized oncogene. Conventional machine learning techniques often necessitate intricate feature engineering and may not capture the full spectrum of sequence dependencies. To ameliorate these limitations, this study employs an adapted U-Net architecture, originally designed for biomedical image segmentation, to classify DNA sequences.The attention mechanism was also tested LONG WITH u-Net architecture to precisely classify DNA sequences into BRAC, BRAF, and KRAS categories. Our comprehensive methodology includes rigorous data preprocessing, model training, and a multi-faceted evaluation approach. The adapted U-Net model exhibited exceptional performance, achieving an overall accuracy of 0.96. The model also achieved high precision and recall rates across the classes, with precision ranging from 0.93 to 1.00 and recall between 0.95 and 0.97 for the key markers BRAC, BRAF, and KRAS. The F1-score for these critical markers ranged from 0.95 to 0.98. These empirical results substantiate the architecture's capability to capture local and global features in DNA sequences, affirming its applicability for critical, sequence-based bioinformatics challenges.

Haplotype-resolved assembly of diploid and polyploid genomes using quantum computing

Precision medicine's emphasis on individual genetic variants highlights the importance of haplotype-resolved assembly, a computational challenge in bioinformatics given its combinatorial nature. While classical algorithms have made strides in addressing this issue, the potential of quantum computing remains largely untapped. Here, we present the vehicle routing problem (VRP) assembler: an approach that transforms this task into a vehicle routing problem, an optimization formulation solvable on a quantum computer. We demonstrate its potential and feasibility through a proof of concept on short synthetic diploid and triploid genomes using a D-Wave quantum annealer. To tackle larger-scale assembly problems, we integrate the VRP assembler with Google's OR-Tools, achieving a haplotype-resolved local assembly across the human major histocompatibility complex (MHC) region. Our results show encouraging performance compared to Hifiasm with phasing accuracy approaching the theoretical limit, underscoring the promising future of quantum computing in bioinformatics.

Abstract 3493: Read-level methylation pattern extraction for high-multiplex large-scale targeted NGS assay

Abstract The abnormal changes in DNA methylation are linked to the early stages of carcinogenesis. Identifying these epigenetic changes in circulating tumor DNA (ctDNA) can reveal potential biomarkers for the early diagnosis of various cancers. However, analyzing such data poses bioinformatics challenges due to the lack of sensitivity in detecting the low abundance ctDNA signals in biopsy samples, which are often overwhelmed by the complexity of libraries containing hundreds of targeted regions. Read-level methylation analysis holds the promise of more in-depth DNA methylation detection due to the wide coverage and high sensitivity of rare signals. However, this approach is hindered by the absence of a standardized workflow capable of generating interpretable reports suitable for both bench scientists and professional bioinformaticians. Here, we present a bioinformatics workflow that examines next-generation sequencing (NGS) data and characterizes the read-level methylation patterns of amplicons. Compared to other currently available tools, our method is designed to work with high-multiplex, large-scale targeted assays. It effectively eliminates the undesired noise derived from sequencing byproducts such as false CpG calls, dimers, and off-target alignments. Additionally, to accommodate the substantial volume of data generated by state-of-the-art NGS platforms, the workflow enables parallel processing of samples compatible with both cloud-based and on-premises computing resources. This workflow provides a comprehensive per-sample visualization of DNA methylation patterns and reports read-level methylation results in a “pattern-as-a-feature” table. In this table, the occurrence of an amplicon epiallelic haplotype (pattern) for every sample is represented as a “feature column” and is aggregated with all patterns discovered in the experiment. These read-level patterns, along with other information, can be used to develop machine learning algorithms to reiteratively harvest true predictive features and penalize confounding signals in predicting cancer diagnosis. Citation Format: Mingda Jin, Masatomo Kaneko, Steven Cen, Hongtao Li, Wei Guo, Xinyi Zhou, Atsuko Fujihara, Tsuyoshi Iwata, Lorenzo Storino Ramacciotti, Divyangi Paralkar, Giovanni E Cacciamani, Manju Aron, Osamu Ukimura, Inderbir S. Gill, Gangning Liang, Andre L. Abreu, Jeffrey Bhasin, Xiaojing Yang, Xi-Yu Jia. Read-level methylation pattern extraction for high-multiplex large-scale targeted NGS assay [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3493.

Sensitive remote homology search by local alignment of small positional embeddings from protein language models

Accurately detecting distant evolutionary relationships between proteins remains an ongoing challenge in bioinformatics. Search methods based on primary sequence struggle to accurately detect homology between sequences with less than 20% amino acid identity. Profile- and structure-based strategies extend sensitive search capabilities into this twilight zone of sequence similarity but require slow pre-processing steps. Recently, whole-protein and positional embeddings from deep neural networks have shown promise for providing sensitive sequence comparison and annotation at long evolutionary distances. Embeddings are generally faster to compute than profiles and predicted structures but still suffer several drawbacks related to the ability of whole-protein embeddings to discriminate domain-level homology, and the database size and search speed of methods using positional embeddings. In this work, we show that low-dimensionality positional embeddings can be used directly in speed-optimized local search algorithms. As a proof of concept, we use the ESM2 3B model to convert primary sequences directly into the 3D interaction (3Di) alphabet or amino acid profiles and use these embeddings as input to the highly optimized Foldseek, HMMER3, and HH-suite search algorithms. Our results suggest that positional embeddings as small as a single byte can provide sufficient information for dramatically improved sensitivity over amino acid sequence searches without sacrificing search speed.

Sensitive remote homology search by local alignment of small positional embeddings from protein language models.

Accurately detecting distant evolutionary relationships between proteins remains an ongoing challenge in bioinformatics. Search methods based on primary sequence struggle to accurately detect homology between sequences with less than 20% amino acid identity. Profile- and structure-based strategies extend sensitive search capabilities into this twilight zone of sequence similarity but require slow pre-processing steps. Recently, whole-protein and positional embeddings from deep neural networks have shown promise for providing sensitive sequence comparison and annotation at long evolutionary distances. Embeddings are generally faster to compute than profiles and predicted structures but still suffer several drawbacks related to the ability of whole-protein embeddings to discriminate domain-level homology, and the database size and search speed of methods using positional embeddings. In this work, we show that low-dimensionality positional embeddings can be used directly in speed-optimized local search algorithms. As a proof of concept, we use the ESM2 3B model to convert primary sequences directly into the 3D interaction (3Di) alphabet or amino acid profiles and use these embeddings as input to the highly optimized Foldseek, HMMER3, and HH-suite search algorithms. Our results suggest that positional embeddings as small as a single byte can provide sufficient information for dramatically improved sensitivity over amino acid sequence searches without sacrificing search speed.

Assessing 2D visual encoding of 3D spatial connectivity.

Introduction: When visualizing complex data, the layout method chosen can greatly affect the ability to identify outliers, spot incorrect modeling assumptions, or recognize unexpected patterns. Additionally, visual layout can play a crucial role in communicating results to peers. Methods: In this paper, we compared the effectiveness of three visual layouts-the adjacency matrix, a half-matrix layout, and a circular layout-for visualizing spatial connectivity data, e.g., contacts derived from chromatin conformation capture experiments. To assess these visual layouts, we conducted a study comprising 150 participants from Amazon's Mechanical Turk, as well as a second expert study comprising 30 biomedical research scientists. Results: The Mechanical Turk study found that the circular layout was the most accurate and intuitive, while the expert study found that the circular and half-matrix layouts were more accurate than the matrix layout. Discussion: We concluded that the circular layout may be a good default choice for visualizing smaller datasets with relatively few spatial contacts, while, for larger datasets, the half- matrix layout may be a better choice. Our results also demonstrated how crowdsourcing methods could be used to determine which visual layouts are best for addressing specific data challenges in bioinformatics.

Using explainable machine learning to uncover the kinase-substrate interaction landscape.

Phosphorylation, a post-translational modification regulated by protein kinase enzymes, plays an essential role in almost all cellular processes. Understanding how each of the nearly 500 human protein kinases selectively phosphorylates their substrates is a foundational challenge in bioinformatics and cell signaling. Although deep learning models have been a popular means to predict kinase-substrate relationships, existing models often lack interpretability and are trained on datasets skewed toward a subset of well-studied kinases. Here we leverage recent peptide library datasets generated to determine substrate specificity profiles of 300 serine/threonine kinases to develop an explainable Transformer model for kinase-peptide interaction prediction. The model, trained solely on primary sequences, achieved state-of-the-art performance. Its unique multitask learning paradigm built within the model enables predictions on virtually any kinase-peptide pair, including predictions on 139 kinases not used in peptide library screens. Furthermore, we employed explainable machine learning methods to elucidate the model's inner workings. Through analysis of learned embeddings at different training stages, we demonstrate that the model employs a unique strategy of substrate prediction considering both substrate motif patterns and kinase evolutionary features. SHapley Additive exPlanation (SHAP) analysis reveals key specificity determining residues in the peptide sequence. Finally, we provide a web interface for predicting kinase-substrate associations for user-defined sequences and a resource for visualizing the learned kinase-substrate associations. All code and data are available at https://github.com/esbgkannan/Phosformer-ST. Web server is available at https://phosformer.netlify.app.

Ascomycetes yeasts: The hidden part of human microbiome.

The fungal component of the microbiota, the mycobiota, has been neglected for a long time due to its poor richness compared to bacteria. Limitations in fungal detection and taxonomic identification arise from using metagenomic approaches, often borrowed from bacteriome analyses. However, the relatively recent discoveries of the ability of fungi to modulate the host immune response and their involvement in human diseases have made mycobiota a fundamental component of the microbial communities inhabiting the human host, deserving some consideration in host-microbe interaction studies and in metagenomics. Here, we reviewed recent data on the identification of yeasts of the Ascomycota phylum across human body districts, focusing on the most representative genera, that is, Saccharomyces and Candida. Then, we explored the key factors involved in shaping the human mycobiota across the lifespan, ranging from host genetics to environment, diet, and lifestyle habits. Finally, we discussed the strengths and weaknesses of culture-dependent and independent methods for mycobiota characterization. Overall, there is still room for some improvements, especially regarding fungal-specific methodological approaches and bioinformatics challenges, which are still critical steps in mycobiota analysis, and to advance our knowledge on the role of the gut mycobiota in human health and disease. This article is categorized under: Immune System Diseases > Genetics/Genomics/Epigenetics Immune System Diseases > Environmental Factors Infectious Diseases > Environmental Factors.

The deep learning applications in IoT-based bio- and medical informatics: a systematic literature review

AbstractNowadays, machine learning (ML) has attained a high level of achievement in many contexts. Considering the significance of ML in medical and bioinformatics owing to its accuracy, many investigators discussed multiple solutions for developing the function of medical and bioinformatics challenges using deep learning (DL) techniques. The importance of DL in Internet of Things (IoT)-based bio- and medical informatics lies in its ability to analyze and interpret large amounts of complex and diverse data in real time, providing insights that can improve healthcare outcomes and increase efficiency in the healthcare industry. Several applications of DL in IoT-based bio- and medical informatics include diagnosis, treatment recommendation, clinical decision support, image analysis, wearable monitoring, and drug discovery. The review aims to comprehensively evaluate and synthesize the existing body of the literature on applying deep learning in the intersection of the IoT with bio- and medical informatics. In this paper, we categorized the most cutting-edge DL solutions for medical and bioinformatics issues into five categories based on the DL technique utilized: convolutional neural network, recurrent neural network, generative adversarial network, multilayer perception, and hybrid methods. A systematic literature review was applied to study each one in terms of effective properties, like the main idea, benefits, drawbacks, methods, simulation environment, and datasets. After that, cutting-edge research on DL approaches and applications for bioinformatics concerns was emphasized. In addition, several challenges that contributed to DL implementation for medical and bioinformatics have been addressed, which are predicted to motivate more studies to develop medical and bioinformatics research progressively. According to the findings, most articles are evaluated using features like accuracy, sensitivity, specificity, F-score, latency, adaptability, and scalability.

Merging and concatenation of sequencing reads: a bioinformatics workflow for the comprehensive profiling of microbiome from amplicon data.

A comprehensive profiling of microbial diversity is essential to understand the ecosystem functions. Universal primer sets such as the 515Y/926R could amplify a part of 16S and 18S rRNA and infer the diversity of prokaryotes and eukaryotes. However, the analyses of mixed sequencing data pose a bioinformatics challenge; the 16S and 18S rRNA sequences need to be separated first and analysed individually/independently due to variations in the amplicon length. This study describes an alternative strategy, a merging and concatenation workflow, to analyse the mixed amplicon data without separating the 16S and 18S rRNA sequences. The workflow was tested with 24 mock community (MC) samples, and the analyses resolved the composition of prokaryotes and eukaryotes adequately. In addition, there was a strong correlation (cor=0.950; P-value=4.754e-10) between the observed and expected abundances in the MC samples, which suggests that the computational approach could infer the microbial proportions accurately. Further, 18 samples collected from the Sundarbans mangrove region were analysed as a case study. The analyses identified Proteobacteria, Bacteroidota, Actinobacteriota, Cyanobacteria, and Crenarchaeota as dominant bacterial phyla and eukaryotic divisions such as Metazoa, Gyrista, Cryptophyta, Chlorophyta, and Dinoflagellata were found to be dominant in the samples. Thus, the results support the applicability of the method in environmental microbiome research. The merging and concatenation workflow presented here requires considerably less computational resources and uses widely/commonly used bioinformatics packages, saving researchers analyses time (for equivalent sample numbers, compared to the conventional approach) required to infer the diversity of major microbial domains from mixed amplicon data at comparable accuracy.