Biology subjects

Young, F.

Publications and source records attributed to Young, F..

8 recordsLinked to original sources

bioRxiv2026

Shared and genus-specific impact of domestication on mammal oral microbiomes revealed through museomics

Animal domestication involves genetic, behavioural and ecological changes, all of which also impact the host-associated microbial communities. Early studies reported diversity loss in the gut microbiomes of domestic and captive animals, but comparisons across a wider diversity of host species showed that loss of microbiota does not always occur. For microbial communities other than the gut, comparisons of multiple domestic-wild pairs are lacking. Here, we use dental calculus metagenomics to investigate the impact of a domesticated lifestyle on the oral microbiome of three pairs of wild and domestic mammals from the genera Equus (plains zebra, horse), Ovis (argali, sheep) and Sus (wild boar, pig). We find that the oral microbiome of the three domesticates differs both in terms of taxonomic composition and functional capacity from their wild counterparts, but the change is smaller in Sus than the other two pairs. While the microbiomes of sheep and horses bear increased resemblance to those of humans compared to their wild relatives, this is unlikely to be due to colonisation by human-derived microbes, as our results suggest that ancestral lineages are mostly retained. Overall, the oral microbiome shows both a consistent signal of domestication, found across all three clades, and a host genus-dependent effect, which likely results from specific changes that underpinned each domestication event.

evolutionary biology↗

bioRxiv2026

iSBEM: An Open-Source Workflow for Automated ROI Targeting in Volume Electron Microscopy

Serial Block Face - Scanning Electron Microscopy (SBF-SEM) is a volume EM method suited to investigate the 3D architecture of tissues and even entire organisms at high resolution. However, imaging large volumes in their entirety is time-consuming and not always necessary. Many research projects have a focused interest in well-defined sub-regions of the samples. The targeting and acquisition of such regions of interest (ROIs) are however currently conducted in a manual way and require heavy involvement of experienced operators. We present a workflow and an original open-source software tool (iSBEM), which allow automated targeting of ROIs in a large tissue sample, based on X-ray microscopy (XRM) maps. After an initial ROI identification and registration of the XRM map with the sample mounted on the SBF-SEM stage, iSBEM takes over the control of the microscope, triggering high resolution acquisitions at defined ROI positions, with minimal user intervention. We demonstrate the approach on two biologically distinct specimens -- malarial oocysts in infected mosquito midgut tissue, and immune cells in human kidney biopsies -- achieving significant improvement in acquisition throughput relative to manual operations, without compromising targeting precision. We also showcase the workflow in a correlative light-Xray-electron microscopy setup, which allowed us to further improve the correct target definition.

cell biology↗

bioRxiv2024

PLM-interact: extending protein language models to predict protein-protein interactions

Computational prediction of protein structure from amino acid sequences alone has been achieved with unprecedented accuracy, yet the prediction of protein-protein interactions (PPIs) remains an outstanding challenge. Here we assess the ability of protein language models (PLMs), routinely applied to protein folding, to be retrained for PPI prediction. Existing PPI prediction models that exploit PLMs use a pre-trained PLM feature set, ignoring that the proteins are physically interacting. Our novel method, PLM-interact, goes beyond a single protein, jointly encoding protein pairs to learn their relationships, analogous to the next-sentence prediction task from natural language processing. This approach provides a significant improvement in performance: Trained on human-human PPIs, PLM-interact predicts mouse, fly, worm, E. coli and yeast PPIs, with 16-28% improvements in AUPR compared with state-of-the-art PPI models. Additionally, it can detect changes that disrupt or cause PPIs and be applied to virus-host PPI prediction. Our work demonstrates that large language models can be extended to learn the intricate relationships among biomolecules from their sequences alone.

bioinformatics↗

bioRxiv2024

From a single sequence to evolutionary trajectories: protein language models capture the evolutionary potential of SARS-CoV-2 protein sequences

Protein language models (PLMs) capture features of protein three-dimensional structure from amino acid sequences alone, without requiring multiple sequence alignments (MSA). The concepts of grammar and semantics from natural language have the potential to capture functional properties of proteins. Here, we investigate how these representations enable assessment of variation due to mutation. Applied to SARS-CoV-2s spike protein using in silico deep mutational scanning (DMS) we demonstrate the PLM, ESM-2, has learned the sequence context within which variation occurs, capturing evolutionary constraint. This recapitulates what conventionally requires MSA data to predict. Unlike other state-of-the-art methods which require protein structures or multiple sequences for training, we show what can be accomplished using an unmodified pretrained PLM. We demonstrate that the grammaticality and semantic scores represent novel metrics. Applied to SARS-CoV-2 variants across the pandemic we show that ESM-2 representations encode the evolutionary history between variants, as well as the distinct nature of variants of concern upon their emergence, associated with shifts in receptor binding and antigenicity. PLM likelihoods can also identify epistatic interactions among sites in the protein. Our results here affirm that PLMs are broadly useful for variant-effect prediction, including unobserved changes, and can be applied to understand novel viral pathogens with the potential to be applied to any protein sequence, pathogen or otherwise.

bioinformatics↗

bioRxiv2023

Prediction of virus-host association using protein language models and multiple instance learning

Predicting virus-host associations is essential to determine the specific host species that viruses interact with, and discover if new viruses infect humans and animals. Currently, the host of the majority of viruses is unknown, particularly in microbiomes. To address this challenge, we introduce EvoMIL, a deep learning method that predicts the host species for viruses from viral sequences only. It also identifies important viral proteins that significantly contribute to host prediction. The method combines a pre-trained large protein language model (ESM) and attention-based multiple instance learning to allow protein-orientated predictions. Our results show that protein embeddings capture stronger predictive signals than sequence composition features, including amino acids, physiochemical properties, and DNA k-mers. In multi-host prediction tasks, EvoMIL achieves median F1 score improvements of 8.6%, 12.3%, and 4.1% in prokaryotic hosts, and 0.5%, 1.8% and 3% in eukaryotic hosts. EvoMIL binary classifiers achieve impressive AUC over 0.95 for all prokaryotic and range from roughly 0.8 to 0.9 for eukaryotic hosts. Furthermore, EvoMIL estimates the importance of single proteins in the prediction task and maps them to an embedding landscape of all viral proteins, where proteins with similar functions are distinctly clustered together, highlighting the ability of EvoMIL to capture key proteins in virus-host specificity. Author summaryBeing able to predict which viruses can infect which host species, and identifying the specific proteins that are involved in these interactions, are fundamental tasks in virology. Traditional methods for predicting these interactions rely on common manual features among proteins, overlooking the structure of the protein "language" encoded in individual proteins. We have developed a novel method that combines a protein language model and multiple instance learning to allow host prediction directly from protein sequences, without the need to extract manual features. This method significantly improved prediction accuracy and revealed key proteins involved in virus-host interactions.

bioinformatics↗

bioRxiv2020

A three-dimensional musculoskeletal model of the dog

Dogs are an interesting object of investigation because of the wide range of body size, body mass, and physique. In the last several years, the number of clinical and biomechanical studies on dog locomotion has increased. However, the relationship between body structure and joint load during locomotion, as well as between joint load and degenerative diseases of the locomotor system (e.g. dysplasia), are not sufficiently understood. In vivo measurements/records of joint forces and loads or deep/small muscles are complex, invasive, and sometimes ethically questionable. The use of detailed musculoskeletal models may help in filling that knowledge gap. We describe here the methods we used to create a detailed musculoskeletal model with 84 degrees of freedom and 134 muscles. Our model has three key-features: Three-dimensionality, scalability, and modularity. We tested the validity of the model by identifying forelimb muscle synergies of a beagle at walk. We used inverse dynamics and static optimization to estimate muscle activations based on experimental data. We identified three muscle synergy groups by using hierarchical clustering. Predicted activation patterns exhibited good agreement with experimental data for most of the forelimb muscles. We expect that our model will speed up the analysis of how body size, physique, agility, and disease influence joint neuronal control and loading in dog locomotion.

bioengineering↗

bioRxiv2019

The nanophthalmos protein TMEM98 inhibits MYRF self-cleavage and is required for eye size specification

The precise control of eye size is essential for normal vision. TMEM98 is a highly conserved and widely expressed gene which appears to be involved in eye size regulation. Mutations in human TMEM98 are found in patients with nanophthalmos (very small eyes) and variants near the gene are associated in population studies with myopia and increased eye size. As complete loss of function mutations in mouse Tmem98 result in perinatal lethality, we produced mice deficient for Tmem98 in the retinal pigment epithelium (RPE), where Tmem98 is highly expressed. These mice have greatly enlarged eyes that are very fragile with very thin retinas. To gain insight into the mechanism of action we used a proximity labelling approach to discover interacting proteins and identified MYRF as an interacting partner. Mutations of MYRF are also associated with nanophthalmos. The protein is an endoplasmic reticulum-tethered transcription factor which undergoes autoproteolytic cleavage to liberate the N-terminal part which then translocates to the nucleus where it acts as a transcription factor. We find that TMEM98 inhibits the self-cleavage of MYRF, in a novel regulatory mechanism. In RPE lacking TMEM98, MYRF is ectopically activated and abnormally localised to the nuclei. Author summaryHaving the correct eye size is important, too large and you will be short-sighted and too small and you will be far-sighted. Nanophthalmos, literally very small eye from the Greek, is a condition where the eye is very small but structurally normal. In addition to being farsighted such eyes are prone to glaucoma which can lead to loss of vision. Here we studied a protein called TMEM98 which is found in the membranes of the cells which form a layer at the back of eye called the retinal pigment epithelium (RPE). Mutations in TMEM98 have been found in nanophthalmos patients. Patients have one normal copy of the gene and one carrying a mutation. We removed Tmem98 from the RPE of mice in order to ascertain its function. We found, surprisingly, that rather than having small eyes this led to the development of very large eyes that were structurally fragile. We went to on to identify protein partners of TMEM98 and found that it interacts with a protein called MYRF, mutations in which also causes nanophthalmos. This work demonstrates the importance of TMEM98 in eye size specification.

genetics↗

bioRxiv2017

Performance of serological antibody tests for bovine tuberculosis in cattle from infected herds in Northern Ireland

The ability to accurately identify infected hosts is the cornerstone of effective disease control and eradication programs. In the case of bovine tuberculosis, caused by infection with the pathogen Mycobacterium bovis, accurately identifying infected individual animals has been challenging as all available tests exhibit less than 100% discriminatory ability. Here we assess the utility of three serological tests and assess their performance relative to skin test (Single Intradermal Comparative Cervical Tuberculin; SICCT), gamma-interferon (IFN{gamma}) and post-mortem results in a Northern Ireland setting. Furthermore, we describe a case-study where one test was used in conjunction with statutory testing.\n\nSerological tests using samples taken prior to SICCT disclosed low proportions of animals as test positive (mean 3% positive), despite the cohort having high proportions with positive SICCT test under standard interpretation (121/921; 13%) or IFN{gamma} (365/922; 40%) results. Furthermore, for animals with a post-mortem record (n=286), there was a high proportion with TB visible lesions (27%) or with laboratory confirmed infection (25%). As a result, apparent sensitivities within this cohort was very low ([≤]15%), however the tests succeeded in achieving very high specificities (96-100%). During the case-study, 7/670 (1.04%) samples from SICCT negative animals from a large chronically infected herd were serology positive, with a further 10 animals being borderline positive (17/670; 2.54%). 9/17 of these animals were voluntarily removed, none of which were found to be infected (-lesions/-bacteriology) post-mortem; 1 serology test negative animal was subsequently lesion+ and M bovis confirmed at slaughter.\n\nImportanceEradication of bovine tuberculosis (bTB; caused by Mycobacterium bovis) has remained elusive in a number of countries despite long-term coordinated test and cull programs. This can partially be explained by the limitations of available statutory tests; therefore supplementary test platforms that identify additional infected animals would be of significant utility. Overall, during our study three serological tests did not disclose a high proportion of animals as infected in high-risk cattle herds, and exhibited limited ability to disclose animals that were positive to the statutory skin test, the gamma interferon test (IFN{gamma}), or were post-mortem confirmed with M. bovis. These serological tests could be used in a supplementary fashion to the statutory tests in particular circumstances; but may be of limited advantage where parallel use of IFN{gamma} and skin testing is performed, as these tests together tended to disclose the majority of animals with post-mortem evidence of infection in our study cohort.

epidemiology↗