The accumulation of senescent cells with age and their pro-inflammatory secretome, known as SASP, contribute to age-related pathologies, including Alzheimer's disease. Therapeutic strategies aimed at eliminating or reprogramming senescent cells require precise knowledge of the genes that drive senescence. However, existing resources such as CellAge, GenAge, SenMayo, and Reactome are built on different operational definitions of senescence and are incompatible with one another. They provide no information on gene conservation in model organisms, tissue expression, or network interactions, all of which are critical for choosing a gene, tissue, and model system for further experiments. SenSASP fills this gap by merging the four sources on a common identifier and adding three layers of annotation missing from all the original resources.
Four gene lists were downloaded from the primary sources: CellAge and GenAge from HAGR, SenMayo from MSigDB, and the Reactome Cellular Senescence pathway via ContentService. Each unique gene symbol was mapped to a canonical set of identifiers (HGNC symbol, Ensembl gene ID, UniProt accession, Entrez ID, HGNC ID) using MyGene.info, with priority given to exact matches and human records. Genes were merged by Ensembl gene ID, with source records kept for each gene. For the conservation layer, mouse and zebrafish orthologs were obtained from Ensembl BioMart. Expression data came from GTEx v8 (median TPM across 54 tissues) and the Human Protein Atlas (specificity, distribution, subcellular localization). High-confidence protein-protein interactions were taken from STRING v12 with a combined score threshold of ≥ 700, and for each gene the network degree and partners were computed.
Merging the four lists yielded 1460 records and 1252 unique symbols, which were mapped to 1250 unique Ensembl genes, removing 210 redundant records (14.4%). Most genes (1077) are specific to a single source, 173 are supported by two or more, 138 by exactly two, 33 by three, and only 2 genes (IL6 and JUN) are backed by all four. Annotations covered nearly all genes: 94.4% have a mouse ortholog, 83.6% have a zebrafish ortholog, 97.9% have GTEx expression profiles and HPA annotations, and 93.0% participate in at least one high-confidence interaction in a network of 15611 edges. The network has a hub-centric topology with TP53 as the main hub (degree 368). 89.4% of genes are annotated across all three layers simultaneously. Clustering of expression profiles for genes supported by three or more sources revealed coherent tissue modules, including broadly expressed regulators and tissue-specific patterns.
SenSASP stands out from neighboring work by merging four heterogeneous resources on aging on a single identifier and adding three layers of annotation absent from any of the original sources. Unlike studies that use large language models or explainable AI methods to analyze genes, SenSASP relies on direct integration of curated data and publicly available APIs, ensuring reproducibility and traceability. It is the first database to link conservation, expression, and interactions for aging genes in one place, turning prioritization and contextualization tasks into queries rather than data integration projects.
The database can be expanded with additional model organisms, cell-type-level expression from single-cell atlases, and directed or regulatory edges in addition to the undirected STRING network. Layers such as GWAS loci or single-cell co-expression could be added for finer prioritization. Users can reweight scoring components or restrict the subset of sources, making SenSASP a flexible platform for hypothesis generation.
The work will influence aging and age-related disease research by providing a ready-made tool for selecting target genes and model systems. It could also accelerate translational research in aging and cancer therapy, where precise knowledge of senescence drivers is required.
Next steps include validating prioritized genes in mouse and zebrafish experiments, as well as integrating single-cell data to refine cellular contexts of expression. It would also be useful to expand the database with additional organisms and regulatory interactions.
The work directly addresses an unresolved problem in aging biology: which genes are true drivers of senescence and SASP, rather than mere correlates. By combining multiple lines of evidence and adding functional annotations, SenSASP helps narrow the pool of candidates for experimental testing, which is critical for developing senolytics and senomorphics.
This section is addressed to the author of the paper and their colleagues. Neighbouring works were found by meaning-based search across our archive, not by word overlap. We do not review: we show the field around the work.
What the work does
The work brings together four gene databases of cellular aging, CellAge, GenAge, SenMayo and Reactome, on a shared Ensembl identifier, adding annotations for conservation in mouse and zebrafish, expression across 54 GTEx tissues, and protein-protein interactions from STRING. The resulting database contains 1250 unique genes, of which 173 are confirmed by two or more sources (IL6 and JUN by all four), and it makes it possible to prioritize candidates and map tissue expression.
Where its strength is
The strength of the work lies in the practical merging of heterogeneous resources on a single identifier with the addition of three layers of functional annotation, which yields a ready-to-use tool for selecting genes, tissues and model organisms in aging experiments.
Where it could go next
Worth trying
Why this may matter
Such extensions could turn SenSASP from a reference resource into a platform for generating hypotheses about the mechanisms of aging and age-related diseases, linking genetic drivers to epigenetic and regulatory processes.
Nearby works in our archive
The archive is sparse around this work: the closest review is only 0.49, against a usual 0.64. It looks like few have drilled nearby.
Only two genes, IL6 and JUN, are supported by all four original resources, highlighting the fragmentation of the field and the importance of a consensus core. TP53, the network's main hub, has degree 368, reflecting its central role in stress-induced senescence.
Not published on arXiv. Here are its page, the PDF and the formalization: you can check it yourself, not only read our account.