Biology subjects

Winther, O.

Publications and source records attributed to Winther, O..

At least 19 recordsLinked to original sources

bioRxiv2026

DeepTMHMM2 enables accurate prediction of transmembrane protein topology and subcellular location

Transmembrane -helical and {beta}-barrel proteins are a ubiquitous component of proteomes. Topology prediction infers how proteins are embedded in lipid bilayers, identifying membrane-spanning segments and their orientation. While recent methods achieve high performance for membrane-spanning segments, they cannot predict re-entrant regions and interfacial helices - membrane-associated segments that partially insert but do not cross the bilayer - nor identify which biological membrane a protein resides in. Here, we present DeepTMHMM2, the first predictor to include re-entrant regions and interfacial helices in its topologies and jointly predict localization across 17 biological membranes. Benchmark results show that DeepTMHMM2 successfully learns to predict the additional elements, while achieving strong performance on canonical -helical and {beta}-barrel topology prediction. Applying DeepTMHMM2 to Swiss-Prot reveals that non-crossing segments are a ubiquitous feature of the transmembrane proteome, with interfacial helices present in nearly a quarter of all -helical transmembrane proteins.

bioinformatics↗

bioRxiv2026

DeepCDS: Ab initio coding sequence prediction in prokaryotic short reads

Accurate coding sequence prediction in short prokaryotic metagenomic sequences, such as Illumina reads, remains challenging due to sequence fragmentation, unknown sequence origins, and sequencing errors. Here we introduce DeepCDS, a deep learning-based ab initio coding sequence predictor trained on short prokaryotic sequences with and without simulated Illumina-like sequencing errors. DeepCDS integrates ESM-2 protein language model embeddings with nucleotide-level information to predict complete and fragmented coding sequence regions. Benchmarking on 215 phylogenetically diverse prokaryotic organisms demonstrates that DeepCDS consistently outperforms current state-of-the-art methods in coding sequence detection across sequence lengths from 75bp to 1000bp, including improved identification of both canonical and non-canonical start codons as well as stop codons, and greater robustness to different sequencing error profiles. DeepCDS also remains operational on sequences shorter than existing tools support. We additionally evaluate DeepCDS on a simulated metagenomic dataset from the Critical Assessment of Metagenome Interpretation initiative, where it likewise achieves the strongest performance among tested tools. These findings demonstrate that protein language models capture distinct signals relevant for nucleotide-level coding sequence detection, especially at very short lengths. Ultimately, DeepCDS may help uncover the functional potential of the vast microbial diversity that remains genomically uncharacterized.

bioinformatics↗

bioRxiv2025

SpatialFormer: Universal Spatial Representation Learning from Subcellular Molecular to Multicellular Landscapes

Spatial transcriptomics quantifies gene expression within its spatial context, significantly advancing biomedical research. Understanding gene spatial expression and the organization of multicellular systems is vital for disease diagnosis and studying biological processes. However, existing models often struggle to integrate gene expression data with cellular spatial information effectively. In this study, we introduce SpatialFormer, a hybrid framework combining convolutional networks and transformers to learn single-cell multimodal and multi-scale information in the niche context, including expression data and subcellular gene spatial distribution. Pre-trained on 300 million cell pairs from 12 million spatially resolved single cells across 62 Xenium slides, SpatialFormer merges gene spatial expression profiles with cell niche information via the pair-wise training strategy. Our findings demonstrate that SpatialFormer distills biological signals across various tasks, including single-cell batch correction, cell-type annotation, co-localization detection, and identifying gene pairs that are critical for the immune cell-cell interactions involved in the regulation of lung fibrosis. These advancements enhance our understanding of cellular dynamics and offer new pathways for applications in biomedical research.

bioinformatics↗

bioRxiv2024

Predicting the subcellular location of prokaryotic proteins with DeepLocPro

Protein subcellular location prediction is a widely explored task in bioinformatics because of its importance in proteomics research. We propose DeepLocPro, an extension to the popular method DeepLoc, tailored specifically to archaeal and bacterial organisms. DeepLocPro is a multiclass subcellular location prediction tool for prokaryotic proteins, trained on experimentally verified data curated from UniProt and PSORTdb. DeepLocPro compares favorably to the PSORTb 3.0 ensemble method, surpassing its performance across multiple metrics on our benchmark experiment. The DeepLocPro prediction tool is available online at https://ku.biolib.com/deeplocpro and https://services.healthtech.dtu.dk/services/DeepLocPro-1.0/.

bioinformatics↗

bioRxiv2023

DeepLocRNA: An Interpretable Deep Learning Model for Predicting RNA Subcellular Localization with domain-specific transfer-learning

Accurate prediction of RNA subcellular localization plays an important role in understanding cellular processes and functions. Although post-transcriptional processes are governed by trans-acting RNA-binding proteins (RBPs) through interaction with cis-regulatory RNA motifs, current methods do not incorporate RBP-binding information. In this paper, we propose DeepLocRNA, an interpretable deep-learning model that leverages a pre-trained multi-task RBP-binding prediction model to predict the subcellular localisation of RNA molecules via fine-tuning. We constructed DeepLocRNA using a comprehensive dataset with variant RNA types and evaluated it on held-out RNA species. Our model achieved state-of-the-art performance in predicting RNA subcellular localization in mRNA and miRNA. It has demonstrated great generalization capabilities, not only for human RNA but also for mice. Moreover, the interpretability of the model is enhanced through the motif analysis, enabling the understanding of the signal factors that contribute to the predictions. The proposed model provides general and powerful prediction abilities for different RNA and species, offering valuable insights into the localisation patterns of RNA molecules and contributing to advancing our understanding of cellular processes at the molecular level.

bioinformatics↗

bioRxiv2023

GraphPart: Homology partitioning for biological sequence analysis

When splitting biological sequence data for the development and testing of predictive models, it is necessary to avoid too closely related pairs of sequences ending up in different partitions. If this is ignored, performance estimates of prediction methods will tend to be exaggerated. Several algorithms have been proposed for homology reduction, where sequences are removed until no too closely related pairs remain. We present GraphPart, an algorithm for homology partitioning, where as many sequences as possible are kept in the dataset, but partitions are defined such that closely related sequences always end up in the same partition. Evaluation of GraphPart on Protein, DNA and RNA datasets shows that it is capable of retaining a larger number of sequences per dataset, while providing homology separation quality on par with reduction approaches.

bioinformatics↗

bioRxiv2023

DiscoTope-3.0 - Improved B-cell epitope prediction using AlphaFold2 modeling and inverse folding latent representations

Accurate computational identification of B-cell epitopes is crucial for the development of vaccines, therapies, and diagnostic tools. However, current structure-based prediction methods face limitations due to the dependency on experimentally solved structures. Here, we introduce DiscoTope-3.0, a markedly improved B-cell epitope prediction tool that innovatively employs inverse folding structure representations and a positive-unlabelled learning strategy, and is explicitly adapted for both solved and predicted structures. Our tool demonstrates a considerable improvement in performance over existing methods, accurately predicting linear and conformational epitopes across multiple independent datasets. Most notably, DiscoTope-3.0 maintains high predictive performance across solved, relaxed and predicted structures, alleviating the need for experimental validation and extending the general applicability of accurate B-cell epitope prediction by more than 3 orders of magnitude. DiscoTope-3.0 is made widely accessible on two web servers, processing over 100 structures per submission, and as a downloadable package. In addition, the servers interface with RCSB and AlphaFoldDB, facilitating large-scale prediction across over 200 million cataloged proteins. DiscoTope-3.0 is available at: https://services.healthtech.dtu.dk/service.php?DiscoTope-3.0

immunology↗

bioRxiv2022

MembraneFold: Visualising transmembrane protein structure and topology

BackgroundAlphaFolds accuracy, which is often comparable to that of experimentally determined structures, has revolutionized protein structure research. Being a statistical method, AlphaFold implicitly infers the cellular environment, e.g. the cell membrane, from the protein sequence. Membrane protein topology prediction methods predict the cellular environment for each protein residue but not the structure. Current structure and topology tools thus provide complementary information. ResultsWe introduce the web server MembraneFold. MembraneFold combines protein structure (from an uploaded PDB file/AlphaFold DB/OmegaFold) and topology (DeepTMHMM) prediction in one server. The output is shown both as a structure with topology superimposed and as a sequence annotation. MembraneFold uses structures predicted by OmegaFold if neither a PDB file is uploaded nor the structure is available in AlphaFold DB. ConclusionMembraneFold is a user-friendly web server that provides practitioners with fast and accurate information about membrane proteins. It is available at https://ku.biolib.com/MembraneFold/.

bioinformatics↗

bioRxiv2022

Identifying endogenous peptide receptors bycombining structure and transmembrane topologyprediction

Many secreted endogenous peptides rely on signalling pathways to exert their function in the body. While peptides can be discovered through high throughput technologies, their cognate receptors typically cannot, hindering the understanding of their mode of action. We investigate the use of AlphaFold-Multimer for identifying the cognate receptors of secreted endogenous peptides in human receptor libraries without any prior knowledge about likely candidates. We find that AlphaFolds predicted confidence metrics have strong performance for prioritizing true peptide-receptor interactions. By applying transmembrane topology prediction using DeepTMHMM, we further improve performance by detecting and filtering biologically implausible predicted interactions. In a library of 1112 human receptors, the method ranks true receptors in the top percentile on average for 11 benchmark peptide-receptor pairs.

bioinformatics↗

bioRxiv2022

Towards In-Silico CLIP-seq: Predicting Protein-RNA Interaction via Sequence-to-Signal Learning

AO_SCPLOWBSTRACTC_SCPLOWUnraveling sequence determinants which drive protein-RNA interaction is crucial for studying binding mechanisms and the impact of genomic variants. While CLIP-seq allows for transcriptome-wide profiling of in vivo protein-RNA interactions, it is limited to expressed transcripts, requiring computational imputation of missing binding information. Existing classification-based methods predict binding with low resolution and depend on prior labeling of transcriptome regions for training. We present RBPNet, a novel deep learning method, which predicts CLIP crosslink count distribution from RNA sequence at single-nucleotide resolution. By training on up to a million regions, RBPNet achieves high generalization on eCLIP, iCLIP and miCLIP assays, outperforming state-of-the-art classifiers. CLIP-seq suffers from various technical biases, complicating downstream interpretation. RBPNet performs bias correction by modeling the raw signal as a mixture of the protein-specific and background signal. Through model interrogation via Integrated Gradients, RBPNet identifies predictive sub-sequences corresponding to known binding motifs and enables variant-impact scoring via in silico mutagenesis. Together, RBPNet improves inference of protein-RNA interaction, as well as mechanistic interpretation of predictions.

bioinformatics↗

bioRxiv2022

Transfer learning reveals sequence determinants of regulatory element accessibility

Dysfunction of regulatory elements through genetic variants is a central mechanism in the pathogenesis of disease. To better understand disease etiology, there is consequently a need to understand how DNA encodes regulatory activity. Deep learning methods show great promise for modeling of biomolecular data from DNA sequence but are limited to large input data for training. Here, we develop ChromTransfer, a transfer learning method that uses a pre-trained, cell-type agnostic model of open chromatin regions as a basis for fine-tuning on regulatory sequences. We demonstrate superior performances with ChromTransfer for learning cell-type specific chromatin accessibility from sequence compared to models not informed by a pre-trained model. Importantly, ChromTransfer enables fine-tuning on small input data with minimal decrease in accuracy. We show that ChromTransfer uses sequence features matching binding site sequences of key transcription factors for prediction. Together, these results demonstrate ChromTransfer as a promising tool for learning the regulatory code.

bioinformatics↗

bioRxiv2022

DeepTMHMM predicts alpha and beta transmembrane proteins using deep neural networks

Transmembrane proteins span the lipid bilayer and are divided into two major structural classes, namely alpha helical and beta barrels. We introduce DeepTMHMM, a deep learning protein language model-based algorithm that can detect and predict the topology of both alpha helical and beta barrels proteins with unprecedented accuracy. DeepTMHMM (https://dtu.biolib.com/DeepTMHMM) scales to proteomes and covers all domains of life, which makes it ideal for metagenomics analyses.

bioinformatics↗

bioRxiv2021

Deep integrative models for large-scale human genomics

Polygenic risk scores (PRSs) are expected to play a critical role in achieving precision medicine. Currently, PRS predictors are generally based on linear models using summary statistics, and more recently individual-level data. However, these predictors mainly capture additive relationships and are limited in data modalities they can use. Here, we developed a deep learning framework (EIR) for PRS prediction which includes a model, genome-local-net (GLN), specifically designed for large scale genomics data. The framework supports multi-task (MT) learning, automatic integration of other clinical and biochemical data, and model explainability. When applied to individual level data in the UK Biobank, we found that GLN outperformed LASSO for a wide range of diseases and in particularly autoimmune diseases. Furthermore, we show that this was likely due to modelling epistasis, and we showcase this by identifying widespread epistasis for Type 1 Diabetes. Furthermore, we trained PRS by integrating genotype, blood, urine and anthropometrics and found that this improved performance for 93% of 290 diseases and disorders considered. Finally, we found that including genotype data provided better calibrated PRS models compared to using measurements alone. EIR is available at https://github.com/arnor-sigurdsson/EIR.

bioinformatics↗

bioRxiv2021

SignalP 6.0 achieves signal peptide prediction across all types using protein language models

Signal peptides (SPs) are short amino acid sequences that control protein secretion and translocation in all living organisms. As experimental characterization of SPs is costly, prediction algorithms are applied to predict them from sequence data. However, existing methods are unable to detect all known types of SPs. We introduce SignalP 6.0, the first model capable of detecting all five SP types. Additionally, the model accurately identifies the positions of regions within SPs, revealing the defining biochemical properties that underlie the function of SPs in vivo. Results show that SignalP 6.0 has improved prediction performance, and is the first model to be applicable to metagenomic data. SignalP 6.0 is available at https://services.healthtech.dtu.dk/service.php?SignalP-6.0

bioinformatics↗

bioRxiv2021

A comparison of tools for copy-number variation detection in germline whole exome and whole genome sequencing data

BackgroundCopy-number variations (CNVs) have important clinical implications for several diseases and cancers. The clinically relevant CNVs are hard to detect because CNVs are common structural variations that define large parts of the normal human genome. CNV calling from short-read sequencing data has the potential to leverage available cohort studies and allow full genomic profiling in the clinic without the need for additional data modalities. Questions regarding performance of CNV calling tools for clinical use and suitable sequencing protocols remain poorly addressed, mainly because of the lack of good reference data sets. MethodsWe reviewed 50 popular CNV calling tools and included 11 tools for benchmarking in a unique reference cohort encompassing 39 whole genome sequencing (WGS) samples paired with analysis by the current clinical standard --SNP-array based CNV calling. Additionally, for nine of these samples we performed whole exome sequencing (WES) performed, in order to address the effect of sequencing protocol on CNV calling. Furthermore, we included Gold Standard reference sample NA12878, and tested 12 samples with CNVs confirmed by multiplex ligation-dependent probe amplification (MLPA). ResultsTool performance varied greatly in the number of called CNVs and bias for CNV lengths. Some tools had near-perfect recall of CNVs from arrays for some samples, but poor precision. Filtering output by CNV ranks from tools did not salvage precision. Several tools had better performance patterns for NA12878, and we hypothesize that this is the result of overfitting during the tool development. ConclusionsWe suggest combining tools with the best recall: GATK gCNV, Lumpy, DELLY, and cn.MOPS. These tools also capture different CNVs. Further improvements in precision requires additional development of tools, reference data sets, and annotation of CNVs, potentially assisted by the use of background panels for filtering of frequently called variants.

bioinformatics↗

bioRxiv2020

COVID-19 Vaccine Candidates: Prediction and Validation of 174 SARS-CoV-2 Epitopes

The recent outbreak of SARS-CoV-2 (2019-nCoV) virus has highlighted the need for fast and efficacious vaccine development. Stimulation of a proper immune response that leads to protection is highly dependent on presentation of epitopes to circulating T-cells via the HLA complex. SARS-CoV-2 is a large RNA virus and testing of all overlapping peptides in vitro to deconvolute an immune response is not feasible. Therefore HLA-binding prediction tools are often used to narrow down the number of peptides to test. We tested 19 epitope-HLA-binding prediction tools, and using an in vitro peptide MHC stability assay, we assessed 777 peptides that were predicted to be good binders across 11 MHC allotypes. In this investigation of potential SARS-CoV-2 epitopes we found that current prediction tools vary in performance when assessing binding stability, and they are highly dependent on the MHC allotype in question. Designing a COVID-19 vaccine where only a few epitope targets are included is therefore a very challenging task. Here, we present 174 SARS-CoV-2 epitopes with high prediction binding scores, validated to bind stably to 11 HLA allotypes. Our findings may contribute to the design of an efficacious vaccine against COVID-19.

immunology↗

bioRxiv2020

Language modelling for biological sequences - curated datasets and baselines

MotivationLanguage modelling (LM) on biological sequences is an emergent topic in the field of bioinformatics. Current research has shown that language modelling of proteins can create context-dependent representations that can be applied to improve performance on different protein prediction tasks. However, little effort has been directed towards analyzing the properties of the datasets used to train language models. Additionally, only the performance of cherry-picked downstream tasks are used to assess the capacity of LMs. ResultsWe analyze the entire UniProt database and investigate the different properties that can bias or hinder the performance of LMs such as homology, domain of origin, quality of the data, and completeness of the sequence. We evaluate n-gram and Recurrent Neural Network (RNN) LMs to assess the impact of these properties on performance. To our knowledge, this is the first protein dataset with an emphasis on language modelling. Our inclusion of properties specific to proteins gives a detailed analysis of how well natural language processing methods work on biological sequences. We find that organism domain and quality of data have an impact on the performance, while the completeness of the proteins has little influence. The RNN based LM can learn to model Bacteria, Eukarya, and Archaea; but struggles with Viruses. By using the LM we can also generate novel proteins that are shown to be similar to real proteins. Availability and implementationhttps://github.com/alrojo/UniLanguage

bioinformatics↗

bioRxiv2019

Detecting Novel Sequence Signals in Targeting Peptides Using Deep Learning

In bioinformatics, machine learning methods have been used to predict features embedded in the sequences. In contrast to what is generally assumed, machine learning approaches can also provide new insights into the underlying biology. Here, we demonstrate this by presenting TargetP 2.0, a novel state of art method to identify N-terminal sorting signals, which direct proteins to the secretory pathway, mitochondria and chloroplasts or other plastids.\n\nBy examining the strongest signals from the attention layer in the network, we find that the second residue in the protein, i.e. the one following the initial methionine, has a strong influence on the classification. When subsequently examining all targeting peptides, we observe that two-thirds of chloroplast and thylakoid transit peptides have an alanine in position two, but only 20% of other plant proteins. Further highlighting the importance of the second residue, we also note that in fungi and single-celled eukaryotes, less than 30% of the targeting peptides have an amino acid that allows the removal of the N-terminal methionine compared with 60% for the proteins without targeting peptide.\n\nTargetP 2.0 is available at http://www.cbs.dtu.dk/services/TargetP-2.0/index.php

bioinformatics↗