This is a dataset describes transcript expression based on the data
from [GTEX](https://gtexportal.org/home/downloads/adult-gtex/bulk_tissue_expression) (
file: GTEx_Analysis_2017-06-05_v8_RSEMv1.3.0_transcript_tpm.gct.gz )

It measures the transcription level of each human transcripts ( 4-5 per genes ) in every
31 tissues (for the V1, 50 more specific tissues for the V2).

expression values are log-transformed ( x -> log(x+1) )

Each sample of the dataset correspond to:

* A unique (gene_id, transcript_id) pair
* the quantified level of expressions across several tissues
* the DNA sequence
* the RNA sequence (ATCG format)
* the protein sequence

## Splits Statistics

|              | train | test | evaluation |
|--------------|-------|------|------------|
| # Genes      | 17648 | 980  | 981        |
| # Transcript | 82205 | 4718 | 4318       |



<img src="./tr_per_gene.png" width="512"/>

<img src="./tr_expr_among_gene.png" width="512"/>

<img src="./rna_length.png" width="512"/>

<img src="./protein_length.png" width="512"/>


### NOTE

We currently only included coding genes, there there the RNA is actually mRNA