Biology subjects

Vijayaraghavan, A.

Publications and source records attributed to Vijayaraghavan, A..

5 recordsLinked to original sources

bioRxiv2025

Microbial Named Entity Recognition and Normalisation for AI-assisted Literature Review and Meta-Analysis

MotivationManual curation of biomedical literature is slow and error-prone and while large language models (LLMs) trained on general texts have shown to be useful for text summarisation, these methods lack the domain-specific expertise required to perform this task accurately. Here we describe the creation of the first microbiome-specific text corpus, use this to train deep learning algorithms for named-entity recognition (NER) and normalisation (NEN), and demonstrate their use to meta-analyse microbiome literature. MethodsWe developed an automated pipeline to annotate all mentions of bacteria, archaea, and fungi in 1,410 full-text microbiome articles. We manually annotated (gold-standard) a separate test set of 288 documents. We trained different transformer-based language models for microbiome recognition and normalisation to taxonomic identifiers and evaluate their performance using the precision, recall, F1-score, and accuracy on the test set. The best models were used to automatically annotate all available Open Access, full-text microbiome articles (n=6,927) and identify taxa that are significantly overrepresented across 14 domains. ResultsThe training and validation set contained a total of 90,150 annotations (both long form and abbreviations). Using the gold-standard test set, with an inter-annotator agreement rate of 99.52% for NER and 88.31% for NEN, the trained models were evaluated and our fine-tuned BioBERT model achieved an F1-score of 96% for NER surpassing a rule- and dictionary-based annotation pipeline (94%). For NEN the accuracy obtained by the deep learning models greatly surpassed that of the pipeline (91% vs 69%). Evaluated across the entire available literature, our models annotate an entire full-text document in only 7 seconds. ConclusionOur algorithms have near perfect precision and greatly speed up the process of annotating microbes in full-text articles. We demonstrated the capabilities of these methods by analysing the entire available literature and describe the taxa associated with each of the domains in our meta-analysis, and exemplify how these methods can be integrated into literature review workflows improving both the speed and accuracy of results. AvailabilityAll codes and data for automatic annotation, model training, and generation of taxonomic trees visualising the data will be made available following peer review with instructions on how to deploy the model on new texts from https://github.com/omicsNLP/microbELP.

bioinformatics↗

bioRxiv2023

Vocabulary Matters: An Annotation Pipeline and Two Deep Learning Algorithms for Enzyme Named Entity Recognition

Enzymes are indispensable substances in many biological processes. With biomedical literature growing exponentially, it becomes more difficult to review the literature effectively. Hence, text-mining techniques are needed to facilitate and speed up literature review. The aims of this study are to create a corpus with annotated enzymes to train and evaluate enzyme named-entity recognition (NER) models. A novel pipeline was built using a combination of dictionary matching and rulebased keyword searching to automatically annotate enzyme entities in over 4,800 biomedical full texts. Two Bidirectional Long Short-Term Memory (BiLSTM) networks using BioBERT and SciBERT as tokeniser and word embedding layers were trained on this corpus and evaluated on a manually annotated test set of 526 full-text publications. The dictionary- and rule-based annotation pipeline achieved an F1-score of 0.863 (precision 0.996, recall 0.762). The SciBERT-BiLSTM model (F1-score 0.965, precision 0.981, recall 0.954) largely out-performed the BioBERT-BiLSTM model (F1-score 0.955, precision 0.981, recall 0.937). This study contributed a novel dictionary- and rule-based automatic pipeline with almost perfect precision which runs in a matter of seconds on a standard laptop. Both deep learning (DL) models achieved state-of-the-art performance (F1>0.95) for enzyme NER, with the SciBERT-based model outperforming the BioBERT-based model in terms of recall, demonstrating the vocabulary used by models matters. The proposed pipeline with the DL models can facilitate more effective enzyme text-mining and information extraction research for literature review and are the first algorithms specifically for enzyme NER. AvailabilityAll codes are available for automatic annotation and model training (including data), with instructions on how to deploy the model on new text, from https://github.com/omicsNLP/enzymeNER.

bioinformatics↗

bioRxiv2022

Longitudinal monitoring of disease burden and response using ctDNA from dried blood spots in xenograft models

Whole genome sequencing (WGS) of circulating tumour DNA (ctDNA) is now a clinically important biomarker for predicting therapy response, disease burden and disease progression. However, the translation of ctDNA monitoring into vital pre-clinical PDX models has not been possible owing to low circulating blood volumes in small rodents. Here, we describe the longitudinal detection and monitoring of ctDNA from minute volumes of blood in PDX mice. We developed a xenograft Tumour Fraction (xTF) metric using shallow WGS of dried blood spots (DBS), and demonstrate its application to quantify disease burden, monitor treatment response and predict disease outcome in a pre-clinical study of PDX mice. Further, we show how our DBS-based ctDNA assay can be used to detect gene-specific copy number changes and examine the copy number landscape over time. Use of sequential DBS ctDNA assays will transform future trial designs in both mice and patients.

cancer biology↗

bioRxiv2017

Simultaneous CRISPR/Cas9-mediated editing of cassava eIF4E isoforms nCBP-1 and nCBP-2 confers elevated resistance to cassava brown streak disease

Cassava brown streak disease (CBSD) is a major constraint on cassava yields in East and Central Africa and threatens production in West Africa. CBSD is caused by two species of positive sense RNA viruses belonging to the family Potyviridae, genus Ipomovirus: Cassava brown streak virus (CBSV) and Ugandan cassava brown streak virus (UCBSV). Diseases caused by the family Potyviridae require the interaction of viral genome-linked protein (VPg) and host eukaryotic translation initiation factor 4E (eIF4E) isoforms. Cassava encodes five eIF4E proteins: eIF4E, eIF(iso)4E-1, eIF(iso)4E-2, novel cap-binding protein-1 (nCBP-1), and nCBP-2. Protein-protein interaction experiments consistently found that VPg proteins associate with cassava nCBPs. CRISPR/Cas9-mediated genome editing was employed to generate ncbp-1, ncbp-2, and ncbp-1/ncbp-2 mutants in cassava cultivar 60444. Challenge with CBSV showed that ncbp-1/ncbp-2 mutants displayed delayed and attenuated CBSD aerial symptoms, as well as reduced severity and incidence of storage root necrosis. Suppressed disease symptoms were correlated with reduced virus titer in storage roots relative to wild-type controls. Our results demonstrate the ability to modify multiple genes simultaneously in cassava to achieve tolerance to CBSD. Future studies will investigate the contribution of remaining eIF4E isoforms on CBSD and translate this knowledge into an optimized strategy for protecting cassava from disease.

plant biology↗

bioRxiv2017

Genomics-enabled analysis of the emergent disease cotton bacterial blight

Cotton bacterial blight (CBB), an important disease of (Gossypium hirsutum) in the early 20th century, had been controlled by resistant germplasm for over half a century. Recently, CBB re-emerged as an agronomic problem in the United States. Here, we report analysis of cotton variety planting statistics that indicate a steady increase in the percentage of susceptible cotton varieties grown each year since 2009. Phylogenetic analysis revealed that strains from the current outbreak cluster with race 18 Xanthomonas citri pv. malvacearum (Xcm) strains.\n\nIllumina based draft genomes were generated for thirteen Xcm isolates and analyzed along with 4 previously published Xcm genomes. These genomes encode 24 conserved and nine variable type three effectors. Strains in the race 18 clade contain 3 to 5 more effectors than other Xcm strains. SMRT sequencing of two geographically and temporally diverse strains of Xcm yielded circular chromosomes and accompanying plasmids. These genomes encode eight and thirteen distinct transcription activator-like effector genes. RNA-sequencing revealed 52 genes induced within two cotton cultivars by both tested Xcm strains. This gene list includes a homeologous pair of genes, with homology to the known susceptibility gene, MLO. In contrast, the two strains of Xcm induce different clade III SWEET sugar transporters. Subsequent genome wide analysis revealed patterns in the overall expression of homeologous gene pairs in cotton after inoculation by Xcm. These data reveal host-pathogen specificity at the genetic level and strategies for future development of resistant cultivars.\n\nAuthor SummaryCotton bacterial blight (CBB), caused by Xanthomonas citri pv. malvacearum (Xcm), significantly limited cotton yields in the early 20th century but has been controlled by classical resistance genes for more than 50 years. In 2011, the pathogen re-emerged with a vengeance. In this study, we compare diverse pathogen isolates and cotton varieties to further understand the virulence mechanisms employed by Xcm and to identify promising resistance strategies. We generate fully contiguous genome assemblies for two diverse Xcm strains and identify pathogen proteins used to modulate host transcription and promote susceptibility. RNA-Sequencing of infected cotton reveals novel putative gene targets for the development of durable Xcm resistance. Together, the data presented reveal contributing factors for CBB re-emergence in the U.S. and highlight several promising routes towards the development of durable resistance including classical resistance genes and potential manipulation of susceptibility targets.

plant biology↗