Search Results (5)

The rapid growth of biomedical literature makes it challenging for researchers to stay current. Integrating knowledge from various sources is crucial for studying complex biological systems. Traditional text-mining methods often have limited accuracy because they don’t capture semantic and contextual nuances. Deep-learning models can be computationally expensive and typically have low interpretability, though efforts in explainable AI aim to mitigate this. Furthermore, transformer-based models have a tendency to produce false or made-up information—a problem known as hallucination—which is especially prevalent in large language models (LLMs). This study proposes a hybrid approach combining text-mining techniques with graph neural networks (GNNs) and fine-tuned large language models (LLMs) to extend biomedical knowledge graphs and interpret predicted edges based on published literature. An LLM is used to validate predictions and provide explanations. Evaluated on a corpus of experimentally confirmed protein interactions, the approach achieved a Matthews correlation coefficient (MCC) of 0.772. Applied to insomnia, the approach identified 25 interactions between 32 human proteins absent in known knowledge bases, including regulatory interactions between MAOA and 5-HT2C, binding between ADAM22 and 14-3-3 proteins, which is implicated in neurological diseases, and a circadian regulatory loop involving RORB and NR1D1. The hybrid GNN-LLM method analyzes biomedical literature efficiency to uncover potential molecular interactions for complex disorders. It can accelerate therapeutic target discovery by focusing expert verification on the most relevant automatically extracted information. Full article

The body of scientific literature continues to grow annually. Over 1.5 million abstracts of biomedical publications were added to the PubMed database in 2021. Therefore, developing cognitive systems that provide a specialized search for information in scientific publications based on subject area ontology and modern artificial intelligence methods is urgently needed. We previously developed a web-based information retrieval system, ANDDigest, designed to search and analyze information in the PubMed database using a customized domain ontology. This paper presents an improved ANDDigest version that uses fine-tuned PubMedBERT classifiers to enhance the quality of short name recognition for molecular-genetics entities in PubMed abstracts on eight biological object types: cell components, diseases, side effects, genes, proteins, pathways, drugs, and metabolites. This approach increased average short name recognition accuracy by 13%. Full article

People with diabetes are more likely to have severe COVID-19 compared to the general population. Moreover, diabetes and COVID-19 demonstrate a certain parallelism in the mechanisms and organ damage. In this work, we applied bioinformatics analysis of associative molecular networks to identify key molecules and pathophysiological processes that determine SARS-CoV-2-induced disorders in patients with diabetes. Using text-mining-based approaches and ANDSystem as a bioinformatics tool, we reconstructed and matched networks related to hyperglycemia, diabetic complications, insulin resistance, and beta cell dysfunction with networks of SARS-CoV-2-targeted proteins. The latter included SARS-CoV-2 entry receptors (ACE2 and DPP4), SARS-CoV-2 entry associated proteases (TMPRSS2, CTSB, and CTSL), and 332 human intracellular proteins interacting with SARS-CoV-2. A number of genes/proteins targeted by SARS-CoV-2 (ACE2, BRD2, COMT, CTSB, CTSL, DNMT1, DPP4, ERP44, F2RL1, GDF15, GPX1, HDAC2, HMOX1, HYOU1, IDE, LOX, NUTF2, PCNT, PLAT, RAB10, RHOA, SCARB1, and SELENOS) were found in the networks of vascular diabetic complications and insulin resistance. According to the Gene Ontology enrichment analysis, the defined molecules are involved in the response to hypoxia, reactive oxygen species metabolism, immune and inflammatory response, regulation of angiogenesis, platelet degranulation, and other processes. The results expand the understanding of the molecular basis of diabetes and COVID-19 comorbidity. Full article

Hypoglycemia has been recognized as a risk factor for diabetic vascular complications and cognitive decline, but the molecular mechanisms of the effect of hypoglycemia on target organs are not fully understood. In this work, gene networks of hypoglycemia and cardiovascular disease, diabetic retinopathy, diabetic nephropathy, diabetic neuropathy, cognitive decline, and Alzheimer’s disease were reconstructed using ANDSystem, a text-mining-based tool. The gene network of hypoglycemia included 141 genes and 2467 interactions. Enrichment analysis of Gene Ontology (GO) biological processes showed that the regulation of insulin secretion, glucose homeostasis, apoptosis, nitric oxide biosynthesis, and cell signaling are significantly enriched for hypoglycemia. Among the network hubs, INS, IL6, LEP, TNF, IL1B, EGFR, and FOS had the highest betweenness centrality, while GPR142, MBOAT4, SLC5A4, IGFBP6, PPY, G6PC1, SLC2A2, GYS2, GCGR, and AQP7 demonstrated the highest cross-talk specificity. Hypoglycemia-related genes were overrepresented in the gene networks of diabetic complications and comorbidity; moreover, 14 genes were mutual for all studied disorders. Eleven GO biological processes (glucose homeostasis, nitric oxide biosynthesis, smooth muscle cell proliferation, ERK1 and ERK2 cascade, etc.) were overrepresented in all reconstructed networks. The obtained results expand our understanding of the molecular mechanisms underlying the deteriorating effects of hypoglycemia in diabetes-associated vascular disease and cognitive dysfunction. Full article

Glucose variability (GV) has been recognized recently as a promoter of complications and therapeutic targets in diabetes. The aim of this study was to reconstruct and analyze gene networks related to GV in diabetes and its complications. For network analysis, we used the ANDSystem that provides automatic network reconstruction and analysis based on text mining. The network of GV consisted of 37 genes/proteins associated with both hyperglycemia and hypoglycemia. Cardiovascular system, pancreas, adipose and muscle tissues, gastrointestinal tract, and kidney were recognized as the loci with the highest expression of GV-related genes. According to Gene Ontology enrichment analysis, these genes are associated with insulin secretion, glucose metabolism, glycogen biosynthesis, gluconeogenesis, MAPK and JAK-STAT cascades, protein kinase B signaling, cell proliferation, nitric oxide biosynthesis, etc. GV-related genes were found to occupy central positions in the networks of diabetes complications (cardiovascular disease, diabetic nephropathy, retinopathy, and neuropathy) and were associated with response to hypoxia. Gene prioritization analysis identified new gene candidates (THBS1, FN1, HSP90AA1, EGFR, MAPK1, STAT3, TP53, EGF, GSK3B, and PTEN) potentially involved in GV. The results expand the understanding of the molecular mechanisms of the GV phenomenon in diabetes and provide molecular markers and therapeutic targets for future research. Full article