TransTEx expression groupings

(human · mouse · cancer)

A browser tool for looking up TransTEx transcript expression-group classifications across three resources — human tissues (GTEx), a mouse body map, and solid-tumor cancers (TCGA solid tumors) — and downloading any filtered slice as CSV.

Available: https://pallavisurana1.github.io/TransTEx_datasets/

What's in it

The data comes from the TransTEx scoring method, which assigns every transcript to one expression group per resource based on how many tissues (or cancers) it is reliably expressed in.

Normal tissues (human GTEx, mouse body map) — five groups:

Class	Meaning
TSp	Tissue-Specific — reliably expressed in a single tissue
TEn	Tissue-Enhanced — expressed in 2 tissues up to 50% of tissues
Wide	Widespread — expressed in more than 50% of tissues
Low	Low or less expression — below the specificity threshold, some expression
Null	No / minimal expression across all tissues

Solid-tumor cancers (TCGA datasets) — the same logic applied across cancer types, giving CanSp, CanEn, CanWide, CanLow, CanNull, plus the two cross-resource biomarker groups: CanHigh (high in cancer, lost in normal — oncogenic-like ) and NorHigh (high in normal, lost in cancer — Tumor suppresor: TSG-like).

Everything is at transcript / isoform level, mapped to genes.

Schema

One row = one transcript × one context (tissue or cancer type):

transcript_id, gene_id, symbol, species, domain, context, class, value

• domain — one of human_tissue, mouse_tissue, cancer
• context — the tissue (e.g. testis, liver) or cancer type (e.g. GBM, OV)
• class — the expression group (TSp/TEn/Wide/Low/Null for tissues; CanSp/CanEn/CanWide/CanLow/CanNull for cancer)
• value — mean expression (TPM/nTPM) for that transcript in that context

Adjust these columns to match your actual grouping tables. If you also carry transcript biotype (protein_coding, lncRNA, …) or the CanHigh/NorHigh flags, add them as extra columns — the app will pick up any column for display and filtering with a small edit.

Data Citations

If you use this dataset or the associated methodologies, please cite the following publications:

🔬 Methods & Frameworks

• TransTEx Method (Human Transcriptome)

  Surana P, Dutta P, Davuluri RV. TransTEx: novel tissue-specificity scoring method for grouping human transcriptome into different expression groups. Bioinformatics 2024;40(8):btae475. 🔗 doi:10.1093/bioinformatics/btae475

• TSProm (Mouse Body-Map Groupings)

  Surana P, Dutta P, Papineni N, Sathian R, Zhou Z, Liu H, Davuluri RV. TSProm: deep learning framework to predict tissue-specific regulatory logic. NAR Genomics and Bioinformatics 2026;8(2):lqag050. 🔗 doi:10.1093/nargab/lqag050

• STPCaT (Solid Tumors Pan-Cancer Transcriptome)

  Surana P, Obusan M, Davuluri RV. Solid Tumors Pan-Cancer Transcriptome: Tissue/Cancer specific expression groups at the isoform level. bioRxiv 2026. 🔗 doi:10.64898/2026.05.04.722705 (CC-BY-NC 4.0)

---

🌐 Software & Databases

• TransTEx Database: Database Link
• TransTEx R Package: GitHub Repository

---

📊 Underlying Source Data

The datasets curated here are derived from the following foundational resources:

Data Type	Source Resource
Human Normal Tissue	GTEx Consortium (Version 8)
Cancer Tissue	TCGA Solid Tumors (via UCSC Xena)
Mouse Tissue	Mouse Body Map (grouped via TSProm)