A Functional Map of the Human Intrinsically Disordered Proteome

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Март 17, 2024

Abstract Intrinsically disordered regions (IDRs) represent at least one-third of the human proteome and defy established structure-function paradigm. Because IDRs often have limited positional sequence conservation, functional classification using standard bioinformatics is generally not possible. Here, we show that evolutionarily conserved molecular features intrinsically (IDR-ome), termed evolutionary signatures, enable prediction IDR functions. Hierarchical clustering IDR-ome based on signatures reveals strong enrichments for frequently studied functions in transcription RNA processing, as well diverse, rarely functions, ranging from sub-cellular localization biomolecular condensates to cellular signaling, transmembrane transport, constitution cytoskeleton. We exploit information encoded within conservation propose annotations every proteome, inspect correlate with different discover co-occurring scale. Further, identify patterns proteins unknown function disease-risk genes conditions such cancer developmental disorders. Our map should be a valuable resource aids discovery new biology.

Язык: Английский

---

SHARK enables sensitive detection of evolutionary homologs and functional analogs in unalignable and disordered sequences

Proceedings of the National Academy of Sciences, Год журнала: 2024, Номер 121(42)

Опубликована: Окт. 9, 2024

Intrinsically disordered regions (IDRs) are structurally flexible protein segments with regulatory functions in multiple contexts, such as the assembly of biomolecular condensates. Since IDRs undergo more rapid evolution than ordered regions, identifying homology poorly conserved remains challenging for state-of-the-art alignment-based methods that rely on position-specific conservation residues. Thus, systematic functional annotation and evolutionary analysis have been limited, despite them comprising ~21% proteins. To accurately assess between unalignable sequences, we developed an alignment-free sequence comparison algorithm, SHARK (Similarity/Homology Assessment by Relating K-mers). We trained SHARK-dive, a machine learning classifier, which achieved superior performance to standard approaches assessing sequences. Furthermore, it correctly identified dissimilar but functionally analogous IDR-replacement experiments reported literature, whereas tools were incapable detecting relationships. SHARK-dive not only predicts similar at proteome-wide scale also identifies cryptic properties motifs drive remote analogy, thereby providing interpretable experimentally verifiable hypotheses determinants underlie acts alternative alignment facilitate universe.

Язык: Английский

---

Two decades of advances in sequence-based prediction of MoRFs, disorder-to-order transitioning binding regions

Expert Review of Proteomics, Год журнала: 2025, Номер unknown

Опубликована: Янв. 9, 2025

Introduction Molecular recognition features (MoRFs) are regions in protein sequences that undergo induced folding upon binding partner molecules. MoRFs common nature and can be predicted from based on their distinctive sequence signatures.

Язык: Английский

---

Decoding intrinsically disordered regions in biomolecular condensates

Fundamental Research, Год журнала: 2025, Номер unknown

Опубликована: Янв. 1, 2025

Язык: Английский

---

Predicting molecular recognition features in protein sequences with MoRFchibi 2.0

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2025, Номер unknown

Опубликована: Фев. 6, 2025

Abstract Identifying sites within intrinsically disordered regions (IDR) that bind to other proteins remains a significant challenge. Molecular Recognition Features (MoRFs) are subset of segments in IDR proteins, undergoing disorder-to-order transition upon binding. This paper introduces MoRFchibi 2.0, specialized prediction tool designed identify the locations MoRFs protein sequences. Our results show 2.0 outperforms all existing MoRF and general predictors protein-binding IDRs, including top-performing models from CAID rounds 1, 2, 3. Remarkably, surpasses utilize AlphaFold data state-of-the-art language models, achieving superior ROC Precision-Recall curves higher success rates. generates output scores using an ensemble convolutional neural network logistic regression followed by reverse Bayes Rule adjust for priors training data. These reflect probabilities normalized data, making them individually interpretable compatible with tools utilizing same scoring framework. Availability https://mc2.msl.ubc.ca/index.xhtml

Язык: Английский

---

Ecological complexity of zoonotic malaria in macaque natural hosts

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Фев. 10, 2025

Macaques are important reservoirs of zoonotic malaria in Southeast Asia. Although cross-sectional surveys have been conducted macaques, little is known about intra-host infection dynamics and host variation susceptibility to these infectious reservoirs. We performed a longitudinal monitoring Plasmodium Hepatocystis infections by microscopy, species-specific polymerase chain reaction (PCR) targeted amplicon deep sequencing (TADS) three long-tailed macaques 20 pig-tailed two districts Narathiwat Province, southern Thailand. In total, 104 macaques' blood samples were obtained during 5 visits with sequential time intervals 9, 4, 7 12 months. Transiently patent low parasite density ( ≤ 1,050 parasites/µL) occurred while PCR TADS diagnosed 45 (43.27%) one or more species parasites, including knowlesi, P. cynomolgi, inui, fieldi, coatneyi, aff. coatneyi sp. macaques. Compared PCR, additionally detected co-infecting 22 48.89%) samples. living close proximity other infected seven free from the 32-month period. Infections for 4 32 months parasites carrying identical complete mitochondrial genome sequences reaffirmed 10 Potentially new transiently over long period course competitive exclusion seemed occur between taxa. Macaques' Duffy phenotypes did not influence differential infections. These results suggest ecological complexity hemoparasite natural malaria. The could affect transmission control disease.

Язык: Английский

---

MultiCycPermea: accurate and interpretable prediction of cyclic peptide permeability using a multimodal image-sequence model

BMC Biology, Год журнала: 2025, Номер 23(1)

Опубликована: Фев. 27, 2025

Cyclic peptides, known for their high binding affinity and low toxicity, show potential as innovative drugs targeting "undruggable" proteins. However, therapeutic efficacy is often hindered by poor membrane permeability. Over the past decade, FDA has approved an average of one macrocyclic peptide drug per year, with romidepsin being only intracellular site. Biological experiments to measure permeability are time-consuming labor-intensive. Rapid assessment cyclic crucial development. In this work, we proposed a novel deep learning model, dubbed MultiCycPermea, predicting MultiCycPermea extracts features from both image information (2D structural information) sequence (1D peptides. Additionally, substructure-constrained feature alignment module align two types features. made leap in predictive accuracy. in-distribution setting CycPeptMPDB dataset, reduced mean squared error (MSE) approximately 44.83% compared latest model Multi_CycGT (0.29 vs 0.16). By leveraging visual analysis tools, can reveal relationship between modification structures permeability, providing insights improve provides effective tool that accurately predicts offering valuable improving This work paves new path application artificial intelligence assisting design membrane-permeable

Язык: Английский

---

Advancements in One-Dimensional Protein Structure Prediction Using Machine Learning and Deep Learning

Computational and Structural Biotechnology Journal, Год журнала: 2025, Номер unknown

Опубликована: Апрель 1, 2025

The accurate prediction of protein structures remains a cornerstone challenge in structural bioinformatics, essential for understanding the intricate relationship between sequence, structure, and function. Recent advancements Machine Learning (ML) Deep (DL) have revolutionized this field, offering innovative approaches to tackle one- dimensional (1D) structure annotations, including secondary solvent accessibility, intrinsic disorder. This review highlights evolution predictive methodologies, from early machine learning models sophisticated deep frameworks that integrate sequence embeddings pretrained language models. Key advancements, such as AlphaFold's transformative impact on rise (PLMs), enabled unprecedented accuracy capturing sequence-structure relationships. Furthermore, we explore role specialized datasets, benchmarking competitions, multimodal integration shaping state-of-the-art By addressing challenges data quality, scalability, interpretability, task-specific optimization, underscores ML, DL, PLMs 1D while providing insights into emerging trends future directions rapidly evolving field.

Язык: Английский

---

Prediction of Protein-Protein Interaction based on Interaction-Specific Learning and Hierarchical Information

Опубликована: Апрель 10, 2025

Background: Prediction of protein–protein interactions (PPIs) is fundamental for identifying drug targets and understanding cellular processes. The rapid growth PPI studies necessitates the development efficient accurate tools automated prediction PPIs. In recent years, several robust deep learning models have been developed found widespread application in proteomics research. Despite these advancements, current computational still face limitations modeling both pairwise hierarchical relationships between proteins. Results: We present HI-PPI, a novel method that integrates representation network interaction-specific protein-protein interaction prediction. HI-PPI extracts information by embedding structural relational into hyperbolic space. A gated then employed to extract features Experiments on multiple benchmark datasets demonstrate outperforms state-of-the-art methods, improves MicroF1 scores 2.62%–7.09% over second-best method. Moreover, offers explicit interpretability organization within network. distance origin computed naturally reflects level Conclusions: Overall, proposed effectively addresses existing methods. By leveraging structure network, significantly enhances accuracy robustness predictions.

Язык: Английский

---

Probabilistic Annotations of Protein Sequences for Intrinsically Disordered Features

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Дек. 21, 2024

Abstract This paper introduces a novel platform for IDR Probabilistic Annotation (IPA). The IPA now encompasses tools predicting ‘Linker’ regions and ‘nucleic’, ‘protein’, ‘all’ (protein or nucleic) binding sites within protein amino acid sequences. Despite its simplicity computational efficiency, results demonstrate that performs competitively with leading in ‘protein’ while considerably outperforming all identifying Linker nucleic sites. An important contribution of this work is the introduction new output paradigm feature predictions. Traditional typically express predictions as scores, higher values indicating greater probabilities. However, these scores lack true probabilistic meaning interpretability, even derived from logistic regression models. limitation arises primarily because training data priors differ broader populations’ unknown priors. proposes applying reverse Bayes Rule to outputs, effectively normalizing data. adjustment produces representing actual probabilities, assuming 50% general population. Such are interpretable isolation enable comparability integration across different tools, marking significant step toward standardization prediction methodologies. Availability orca.msl.ubc.ca/nmshare/ipa.tar.gz

Язык: Английский

---