Research indexMeta-analysis
Бенчмаркинг генных сигнатур и биоинформатических инструментов для оценки сенесценции с использованием данных транскриптома единичных клеток
10 Aug 2026Open access
NobleIDark:/48914/w1/20261105/77A44D05DOI10.18699/bgrs2026-1705
Бенчмаркинг генных сигнатур и биоинформатических инструментов для оценки сенесценции с использованием данных транскриптома единичных клеток
2026
NobleIDark:/48914/w1/20261105/77A44D05DOI10.18699/bgrs2026-1705
Free to read at the publisher
You can read the abstract here, or the full text at the publisher.
Abstract
и я Тезисы докладовIn this model, senescence induced by DNA double-strand breaks represents a key mechanism driving lung fibrosis, with alveolar type II epithelial cells identified as the most susceptible cell population initiating the profibrotic cascade.According to the results obtained, the highest total scores were achieved by methods based on age-related gene sets, specifically gene signatures derived from meta-analyses of age-associated transcriptomic alterations [4,5], as well as the SenePy method using a universal gene signature.The best methods, ranked by total score, demonstrated strong agreement across all evaluated metrics, consistently showing high values for each component of the total score.In contrast, some machine learning-based methods (e.g., SenCID and DeepScence) received lower total scores and did not outperform gene signature-based methods in any of the evaluated metrics.Conclusion: We developed an approach to benchmark senescence evaluation methods using in vivo scRNA-seq data that closely reflect real-world scenarios of senescent cell identification in the absence of ground truth annotations.In the bleomycin-induced lung injury model, methods based on age-associated gene signatures demonstrated the highest sensitivity and specificity, as well as the strongest agreement with experimentally observed senescence dynamics.Machine learning-based approaches did not outperform the other evaluated methods.