Biology subjects

Dunne, M. P.

Publications and source records attributed to Dunne, M. P..

5 recordsLinked to original sources

bioRxiv2026

LLPSight: enhancing prediction of LLPS-driving proteins using machine learning and protein Language Models

In eukaryotic cells, essential functions are often confined within organelles enclosed by lipid membranes. Increasing evidence, however, highlights the role of membrane-less organelles (MLOs), formed through liquid-liquid phase separation (LLPS). MLO assemblies are typically initiated by "driver" proteins, which form a scaffold to recruit additional "client" molecules. By leveraging expanding MLO datasets and modern machine learning approaches, we developed LLPSight, an ML-based predictor of LLPS-driving proteins. The model was trained using rigorously curated datasets: a positive set of proteins experimentally confirmed to drive LLPS in vivo and a negative set of soluble, unstructured proteins not associated with LLPS. For the features, we employed a cutting-edge approach using embeddings from protein Language Models. LLPSight achieves the highest F1 score (0.885) among existing tools, enabling more efficient discovery of new LLPS drivers eagerly awaited by researchers for experimental validation. An additional key feature of LLPSight is its ability to perform proteome-wide analyses; application to the human proteome yielded promising targets. LLPSight can be obtained from authors upon request.

bioinformatics↗

bioRxiv2024

Developing machine-learning-based amyloid predictors with Cross-Beta DB

Due to shifts in environmental conditions, mutations, or interactions with other biomolecules, some proteins that would normally be soluble can undergo aggregation, resulting in the formation of clumps of amyloid fibrils. Understanding of this phenomenon is of paramount importance due not only to its association with various diseases (including Alzheimers disease), but also due to increasingly abundant evidence for its functional roles. Numerous studies have demonstrated that the propensity to form amyloids is coded by the amino acid sequence and this finding has paved the way for the development of several computational predictors of amyloidogenicity. The ultimate objective of computational methods is to accurately predict the formation of disease-related and functionally relevant amyloids that occur in vivo. These amyloid fibrils are known to form very specific "cross-{beta}" structures of protein regions longer than about 15 residues. Remarkably, despite the significance of the naturally occurring amyloids, there has been a lack of datasets specifically dedicated to them. Hence, we built Cross-Beta DB, a database composed of cross-{beta} amyloids formed in natural conditions. This database is expected to be indispensable for benchmarking amyloid predictors. We used the Cross-Beta DB to train and benchmark several such algorithms, using machine learning. The best-performing of these, the random-forest-based Cross-Beta RF Predictor, demonstrated superior performance over the other existing methods, fostering high expectations for an improved prediction of naturally occurring amyloids.

bioinformatics↗

bioRxiv2022

ExpressUrself: A spatial model for predicting recombinant expression from mRNA sequence

Maximising the yield of recombinantly expressed proteins is a critical part of any protein engineering pipeline. In most cases, the expression of a given protein can be tuned by adjusting its DNA coding sequence, however finding coding sequences that optimise expression is a nontrivial task. The 3-dimensional structure of mRNA is known to strongly influence the expression levels of proteins, due to its effect on the efficiency of ribosome attachment. While correlations between mRNA structure and expression are well established, no model to date has succeeded in effectively utilising this information to accurately predict expression levels. Here we present ExpressUrself, a model designed to capture spatial characteristics of the sequence surrounding the start codon of an mRNA transcript, and intended to be used for optimising protein expression. The model is trained and tested on a large data set of variant DNA sequences and is able to predict the expression of previously unseen transcripts to a high degree of accuracy.

bioengineering↗

bioRxiv2021

Deep learning model for the prediction and classification of protein toxins across all domains of life

Toxins are widely produced by different organisms to disrupt the physiology of other organisms, and support their own existence. Their study is useful to understand protein evolution, environmental adaptation and survival competition. In-silico predictions of toxic proteins can support empirical frameworks, and help in the safety measurements needed for various industrial related processes. Some in-silico methods are slow, hard to implement or lack taxa representation in their training datasets. Here we present a deep learning model to classify protein toxins, through the use of Convolutional Neural Networks (ConvTOX). ConvTOX is able to accurately identify toxic proteins across the domains of life, with accuracies over 80% for animal and plant toxins, and over 50% for bacterial toxins. Moreover, ConvTOX is able to generalize the identification of differences among toxin types, such as neurotoxins and myotoxins, and to accurately identify structural similarities between different protein toxins. ConvTOX overcomes limitations from previous models by being able to predict toxin proteins from across all domains of life, and by not being limited to only short toxin peptides. Limitations are still clear in terms of lower accuracies for specific phylogenetic groups (such as bacterial toxins), but still this works presents itself as a one step forward for the universal use, classification and study of toxic proteins.

synthetic biology↗

bioRxiv2017

OrthoFiller: utilising data from multiple species to improve the completeness of genome annotations.

BackroundComplete and accurate annotation of sequenced genomes is of paramount importance to their utility and analysis. Differences in gene prediction pipelines mean that genome sequences for a species can differ considerably in the quality and quantity of their predicted genes. Furthermore, genes that are present in genome sequences sometimes fail to be detected by computational gene prediction methods. Erroneously unannotated genes can lead to oversights and inaccurate assertions in biological investigations, especially for smaller-scale genome projects which rely heavily on computational prediction.\n\nResultsHere we present OrthoFiller, a tool designed to address the problem of finding and adding such missing genes to genome annotations. OrthoFiller leverages information from multiple related species to identify those genes whose existence can be verified through comparison with known gene families, but which have not been predicted. By simulating missing gene annotations in real sequence datasets from both plants and fungi we demonstrate the accuracy and utility of OrthoFiller for finding missing genes and improving genome annotation. Furthermore, we show that applying OrthoFiller to existing \"complete\" genome annotations can identify and correct substantial numbers of erroneously missing genes in these two sets of species.\n\nConclusionsWe show that significant improvements in the completeness of genome annotations can be made by leveraging information from multiple species.

bioinformatics↗