SignatureSets provides access to a curated, literature-backed collection of RNA-based gene expression signatures focused on immuno-oncology (IO) and the tumor microenvironment (TME).
All signatures in the package are:
- Published in peer-reviewed literature
- Publicly available through trusted bioinformatics resources
References for each signature including source publications, accession details, and metadata, are included in:
The repository includes:
- 55 IO (Immuno-Oncology) gene signatures
- 90 TME (Tumor Microenvironment) gene signatures
- TME signatures were extracted using the
IOBRpackage.
- TME signatures were extracted using the
All genes have been standardized to GENCODE v40 annotations using HUGO Gene Symbols, and mapped to Entrez and Ensembl gene identifiers using the biomaRt R package.
SignatureSets/
├── 📁 data/ # Processed signature data (.rda files)
├── 📁 data-raw/ # Processed signature data and metadata (.csv files)
├── 📁 vignettes/ # Summaries and usage examples of signature data
└── 📄 README.md # Repository overview and documentation
The 55 IO signatures are categorized by their association with response to immuno-oncology therapy:
-
36 signatures (65%): Associated with sensitivity to IO therapy, indicating potential positive responses such as immune activation or enhanced checkpoint inhibitor efficacy.
-
19 signatures (35%): Associated with resistance to IO therapy, highlighting mechanisms like immune evasion, suppressive tumor microenvironments, or other resistance pathways.
Signature scores are computed using standardized methods tailored to the characteristics of each signature, as described in their original publications.
-
Unweighted Signatures: Scores are computed using Gene Set Variation Analysis (GSVA) or Single Sample Gene Set Enrichment Analysis(ssGSEA) to assess pathway enrichment. GSVA calculates enrichment scores for gene sets without weighting individual genes.
-
Weighted Signatures: Scores are computed as a weighted mean expression, where weights are assigned as follows: +1 for increased expression and -1 for decreased expression.
-
Specific Algorithm: Certain signature scores are computed based on their respective original publications, e.g., the PredictIO signature.
More details about signature score computations and their applications can be found on the PredictioR GitHub repository.
The package is not yet available on CRAN or Bioconductor. You can install it by cloning the repository:
git clone https://github.com/bhklab/SignatureSets
cd SignatureSets
If you use SignatureSets in your research, please cite:
- Bareche, Y., Kelly, D., Abbas-Aghababazadeh, F. et al., Annals of Oncology 2022.
- For TME signatures extracted via IOBR, please cite the corresponding IOBR publication PMID: 39626665.