Francesca Young

Ms. Francesca Young

MRC-University of Glasgow Centre for Virus Research
orcid

Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
26 May 2020 in PLOS Computational Biology
Verified
Journal Article
Open access
PDF file
FYFrancesca Young
Francesca Young
SRSimon Rogers
Simon Rogers
DRDavid L. Robertson
David L. Robertson
The rise in metagenomics has led to an exponential growth in virus discovery. However, the majority of these new virus sequences have no assigned host. Current machine learning approaches to predicting virus host interactions have a tendency to focus on nucleotide features, ignoring other representations of genomic information. Here we investigate the predictive potential of features generated from four different ‘levels’ of viral genome representation: nucleotide, amino acid, amino acid properties and protein domains. This more fully exploits the biological information present in the virus genomes. Over a hundred and eighty binary datasets for infecting versus non-infecting viruses at all taxonomic ranks of both eukaryote and prokaryote hosts were compiled. The viral genomes were converted into the four different levels of genome representation and twenty feature sets were generated by extracting k-mer compositions and predicted protein domains. We trained and tested Support Vector Machine, SVM, classifiers to compare the predictive capacity of each of these feature sets for each dataset. Our results show that all levels of genome representation are consistently predictive of host taxonomy and that prediction k-mer composition improves with increasing k-mer length for all k-mer based features. Using a phylogenetically aware holdout method, we demonstrate that the predictive feature sets contain signals reflecting both the evolutionary relationship between the viruses infecting related hosts, and host-mimicry. Our results demonstrate that incorporating a range of complementary features, generated purely from virus genome sequences, leads to improved accuracy for a range of virus host prediction tasks enabling computational assignment of host taxonomic information. Elucidating the host of a newly identified virus species is an important challenge, with applications from knowing the source species of a newly emerged pathogen to understanding the bacteriophage-host relationships within the microbiome of any of earth’s ecosystems. Current high throughput methods used to identify viruses within biological or environmental samples have resulted in an unprecedented increase in virus discovery. However, for the majority of these virus genomes the host species/taxonomic classification remains unknown. To address this gap in our knowledge there is a need for fast, accurate computational methods for the assignment of putative host taxonomic information. Machine learning is an ideal approach but to maximise predictive accuracy the viral genomes need to be represented in a format (sets of features) that makes the discriminative information available to the machine learning algorithm. Here, we compare different types of features derived from the same viral genomes for their ability to predict host information. Our results demonstrate that all these feature sets are predictive of host taxonomy and when combined have the potential to improve accuracy over the use of individual feature sets across many virus host prediction applications.

5 July 2017 in Analytical Chemistry
Verified
Journal Article
Open access
Justin Van der HooftJustin Van der Hooft
Justin Van der Hooft
JWJoe Wandy
Joe Wandy
FYFrancesca Young
Francesca Young
Sandosh PadmanabhanSandosh Padmanabhan
Sandosh Padmanabhan
KGKonstantinos Gerasimidis
Konstantinos Gerasimidis
SRSimon Rogers
Simon Rogers
In untargeted metabolomics approaches, the inability to structurally annotate relevant features and map them to biochemical pathways is hampering the full exploitation of many metabolomics experiments. Furthermore, variable metabolic content across samples result in sparse feature matrices that are statistically hard to handle. Here, we introduce MS2LDA+ that tackles both above-mentioned problems. Previously, we presented MS2LDA, which extracts biochemically relevant molecular substructures (“Mass2Motifs”) from a collection of fragmentation spectra as sets of co-occurring molecular fragments and neutral losses, thereby recognizing building blocks of metabolomics. Here, we extend MS2LDA to handle multiple metabolomics experiments in one analysis, resulting in MS2LDA+. By linking Mass2Motifs across samples, we expose the variability in prevalence of structurally related metabolite families. We validate the differential prevalence of substructures between two distinct samples groups and apply it to fecal samples. Subsequently, within one sample group of urines, we rank the Mass2Motifs based on their variance to assess whether xenobiotic-derived substructures are among the most-variant Mass2Motifs. Indeed, we could ascribe 22 out of the 30 most-variant Mass2Motifs to xenobiotic-derived substructures including paracetamol/acetaminophen mercapturate and dimethylpyrogallol. In total, we structurally characterized 101 Mass2Motifs with biochemically or chemically relevant substructures. Finally, we combined the discovered metabolite families with full scan feature intensity information to obtain insight into core metabolites present in most samples and rare metabolites present in small subsets now linked through their common substructures. We conclude that by biochemical grouping of metabolites across samples MS2LDA+ aids in structural annotation of metabolites and guides prioritization of analysis by using Mass2Motif prevalence.

7 November 2024 in bioRxiv
Verified
Preprint
Open access
PDF file
DLDan Liu
Dan Liu
FYFrancesca Young
Francesca Young
KLKieran D. Lamb
Kieran D. Lamb
AQAdalberto Claudio Quiros
Adalberto Claudio Quiros
APAlexandrina Pancheva
Alexandrina Pancheva
KYKe Yuan
Ke Yuan
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.

6 July 2024 in bioRxiv
Verified
Preprint
Open access
PDF file
KLKieran D Lamb
Kieran D Lamb
JHJoseph Hughes
Joseph Hughes
SLSpyros Lytras
Spyros Lytras
FYFrancesca Young
Francesca Young
OKOrges Koci
Orges Koci
DRDavid L Robertson
David L Robertson
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-2's 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.

6 December 2020 in Viruses
Verified
Journal Article
Open access
PDF file
FHFranziska Hufsky
Franziska Hufsky
NBNiko Beerenwinkel
Niko Beerenwinkel
IMIrmtraud M. Meyer
Irmtraud M. Meyer
SRSimon Roux
Simon Roux
GCGeorgia May Cook
Georgia May Cook
VTVolker Thiel
Volker Thiel
The International Virus Bioinformatics Meeting 2020 was originally planned to take place in Bern, Switzerland, in March 2020. However, the COVID-19 pandemic put a spoke in the wheel of almost all conferences to be held in 2020. After moving the conference to 8–9 October 2020, we got hit by the second wave and finally decided at short notice to go fully online. On the other hand, the pandemic has made us even more aware of the importance of accelerating research in viral bioinformatics. Advances in bioinformatics have led to improved approaches to investigate viral infections and outbreaks. The International Virus Bioinformatics Meeting 2020 has attracted approximately 120 experts in virology and bioinformatics from all over the world to join the two-day virtual meeting. Despite concerns being raised that virtual meetings lack possibilities for face-to-face discussion, the participants from this small community created a highly interactive scientific environment, engaging in lively and inspiring discussions and suggesting new research directions and questions. The meeting featured five invited and twelve contributed talks, on the four main topics: (1) proteome and RNAome of RNA viruses, (2) viral metagenomics and ecology, (3) virus evolution and classification and (4) viral infections and immunology. Further, the meeting featured 20 oral poster presentations, all of which focused on specific areas of virus bioinformatics. This report summarizes the main research findings and highlights presented at the meeting.