Biology subjects

Yılmaz, A.

Publications and source records attributed to Yılmaz, A..

3 recordsLinked to original sources

bioRxiv2026

Beyond Chemical Similarity: Structure-Agnostic Drug-Drug Interaction Prediction with MeSH Semantics and a Drug-Target-Protein Knowledge Graph

BackgroundAdverse drug-drug interactions (DDIs) cause preventable hospitalizations, but exhaustive experimental screening of all drug pairs is infeasible. Many computational predictors rely on SMILES or other molecular representations, limiting their direct applicability to biologics and other non-small-molecule therapeutics. We present a structure-agnostic framework that combines semantic representations derived from Medical Subject Headings (MeSH) with graph-derived topology from a Drug-Target-Protein knowledge graph constructed from DrugBank and UniProt. We further investigate how variation in MeSH annotation depth affects predictive performance. ResultsDrugs are grouped according to their deepest MeSH annotation level (Low, Mid, or Deep), and performance is evaluated across the resulting interaction categories in transductive and inductive settings. The Intermediate ontology scope (Low+Mid) provides the most stable performance, while adding Deep-level terms offers limited and inconsistent benefit. Lightweight topological descriptors are integrated with MeSH features through instance-wise, dimension-specific latent-space gating, using curated reliable-negative pairs for supervision. Fusion improves mean performance over the MeSH-only baseline across all six categories in the transductive setting. Under induction, the clearest gains occur for Low-Low interactions ({Delta}AUROC = 0.056;{Delta} F1 = 0.137) and Low-Mid interactions ({Delta}AUROC = 0.077;{Delta} F1 = 0.114). ConclusionsMeSH annotation depth is associated with systematic variation in DDI prediction performance that aggregate evaluation can obscure. Graph-derived topology is particularly beneficial when ontology annotations are shallow. The framework provides a common, structure-agnostic representation compatible with both small-molecule and biologic therapeutics and supports first-pass DDI prioritization for subsequent expert assessment.

bioinformatics↗

bioRxiv2025

Impacts of high pathogenicity avian influenza H5N1 2.3.4.4b south of the Antarctic Circle

High pathogenicity avian influenza (HPAI) H5N1 2.3.4.4b poses a substantial conservation threat to ecosystems, populations, and species globally, with its continued spread into new regions increasing concern for potential ecological consequences. During surveys in February - March 2025, we confirmed the viruss presence at the southern extent of its known range along the Western Antarctic Peninsula, with recorded mortalities in South Polar Skuas Stercorarius maccormicki on distinct islands in Marguerite Bay, as well as one confirmed and one suspected case in Kelp Gulls Larus dominicanus. At the time of sampling, no evidence of infection was observed in other seabird or mammal species. Consistent with previous global reports, skuas - here, South Polar Skuas - appear particularly vulnerable, yet broader impacts on the local seabird and mammal community remain unclear. Additionally, our use of rapid antigen tests (VDRG(R) AIV Ag Rapid kit 2.0 - Median Diagnostics) in the field demonstrated their potential utility for real-time surveillance, though false negatives (10%) highlight limitations in test sensitivity. These findings contribute to a growing understanding of the impacts of HPAI H5N1 2.3.4.4b outbreaks on Antarctic species and populations, and will inform continued monitoring, conservation strategies, and biosecurity measures in response to the viruss ongoing spread.

ecology↗

bioRxiv2022

MTaxi : A comparative tool for taxon identification of ultra low coverage ancient genomes

A major challenge in zooarchaeology is to morphologically distinguish closely related species remains, especially using small bone fragments. Shotgun sequencing aDNA from archeological remains and comparative alignment to the candidate species reference genomes will only apply when reference nuclear genomes of comparable quality are available, and may still fail when coverages are low. Here, we propose an alternative method, MTaxi, that uses highly accessible mitochondrial DNA (mtDNA) to distinguish between pairs of closely related species from ancient DNA sequences. MTaxi utilises mtDNA transversion-type substitutions between pairs of candidate species, assigns reads to either species, and performs a binomial test to determine the sample taxon. We tested MTaxi on sheep/goat and horse/donkey data, between which zooarchaeological classification can be challenging in ways that epitomise our case. The method performed efficiently on simulated ancient genomes down to 0.5x mitochondrial coverage for both sheep/goat and horse/donkey, with no false positives. Trials on n=18 ancient sheep/goat samples and n=10 horse/donkey samples of known species identity with mtDNA coverages 0.1x - 12x also yielded 100% accuracy. Overall, MTaxi provides a straightforward approach to classify closely related species that are compelling to distinguish through zooarchaeological methods using low coverage aDNA data, especially when similar quality reference genomes are unavailable. MTaxi is freely available at https://github.com/goztag/MTaxi.

bioinformatics↗