Transcriptome-Small Molecule Interaction Network (TSI-Net)

Transcriptome-small molecule interaction network (TSI-Net) data, based on the paper "Construction of Multi-Modal Transcriptome-Small Molecule Interaction Networks from High-Throughput Measurements to Study Human Complex Traits" (doi: https://doi.org/10.1101/2025.01.22.634403).

*Please cite the accompanying publication when using these datasets*

The dataset includes 5 files:

"LLFS_TSINet_modules.tsv": This file contains TSI-Net modules inferred from non-diabetic and non-cancer participants of the Long Life Family Study cohort (U19AG063893). Each row represents a single TSI-Net module. The first column contains the module identifier, and from the second column onwards, the gene and small molecule IDs included in the module are listed. There are 388 multi-modal modules in this file.

"LLFS_TSINet_interactions.csv": This file contains gene-small molecule interactions in the LLFS TSI-Net. It includes the following columns:

"gene_id": Ensemble gene ids

"small_molecule": Small molecule names

"module": Module identifier

"source": Indicates the origin of the small molecule profiles. The source is either "polars", representing polar metabolomics profiles, or "lipids", representing lipidomics profiles.

"pval": P-value of the gene-small molecule associations

"Beta" Regression coefficient of the gene-small molecule associations

"knowledge_guided_TSINet_modules.tsv": This file contains TSI-Net modules inferred based on established gene-level modules from the 2019 DREAM Challenge (https://doi.org/10.1038/s41592-019-0509-5). The original sources of these gene-level modules include the STRINGdb protein–protein interaction (PPI) networks, InWeb PPI networks, and GEO co-expression networks. These gene-level modules were integrated with small molecule profiles from the LLFS cohort to construct the knowledge-guided TSI-Net modules. The first column contains the module identifier, and from the second column onwards, the gene and small molecule IDs included in the module are listed. There are 563 multi-modal modules in this file.

"knowledge_guided_TSINet_interactions.csv": This file contains gene-small molecule interactions in the knowledge-guided TSI-Nets. It includes the following columns:

"gene_id": Ensemble gene ids

"small_molecule": Small molecule names

"module": Module identifier

"source": Indicates the origin of the small molecule profiles. The source is either "polars", representing polar metabolomics profiles, or "lipids", representing lipidomics profiles.

"pval": P-value of the gene-small molecule associations

"Beta" Regression coefficient of the gene-small molecule associations

"polars_cid.csv": This metadata file includes the PubChem IDs (CIDs) for the polar metabolomics profiles. It contains two columns:

"small_molecule": The names of small molecules as they appear in the TSI-Net module and interaction files

"CID": PubChem ID