Number of protein-coding transcripts from the gene as defined by Ensembl.

Main subcellular location based on data generated in the subcellular section of the Human Protein Atlas.

All transcripts of all genes have been analyzed regarding the location(s) of corresponding protein based on prediction methods for signal peptides and transmembrane regions.

• Genes with at least one transcript predicted to encode a secreted protein, according to prediction methods or to UniProt location data, have been further annotated and classified with the aim to determine if the corresponding protein(s) are secreted or actually retained in intracellular locations or membrane-attached.


• Remaining genes, with no transcript predicted to encode a secreted protein, will be assigned the prediction-based location(s).

The annotated location overrules the predicted location, so that a gene encoding a predicted secreted protein that has been annotated as intracellular will have intracellular as the final location.

All genes with at least one isoform expected to be secreted to the extracellular environment have been annotated and classified either as secreted to blood or as locally secreted, depending on the predicted final location of the corresponding protein. Proteins expected to be locally secreted have been further classified according to their site of expression.

The RNA data was used to cluster genes according to their expression across tissues. Clusters contain genes that have similar expression patterns, and each cluster has been manually annotated to describe common features in terms of function and specificity.

The regional specificity category is based on mRNA expression levels in the analysed brain samples, grouped into 13 main brain regions and calculated for the three different species. All brain expression profiles are based on data from HPA. The specificity categories include: regionally enriched, group enriched, regionally enhanced, low regional specificity and not detected. The classification rules are the same used for the tissue specificity category

The RNA data was used to cluster genes according to their expression across tissues. Clusters contain genes that have similar expression patterns, and each cluster has been manually annotated to describe common features in terms of function and specificity.

The RNA specificity category is based on mRNA expression levels in the analyzed cell types based on scRNA-seq data from normal tissues. The categories include: cell type enriched, group enriched, cell type enhanced, low cell type specificity and not detected.

The RNA data was used to cluster genes according to their expression across single cell types. Clusters contain genes that have similar expression patterns, and each cluster has been manually annotated to describe common features in terms of function and specificity.

Genes can have enriched specificity in different cell types in one or several tissues, or be enriched in a core cell type that appears in many different tissues.

Specificity of RNA expression in 17 cancer types is categorized as either cancer enriched, group enriched, cancer enhanced, low cancer specificity and not detected.

The RNA data was used to cluster genes according to their expression across cell lines. Clusters contain genes that have similar expression patterns, and each cluster has been manually annotated to describe common features in terms of function and specificity.

RNA specificity category based on RNA sequencing data from cancer cell lines in the Human Protein Atlas grouped according to type of cancer. Genes are classified into six different categories (enriched, group enriched, enhanced, low specificity and not detected) according to their RNA expression levels across the panel of cell lines.

A gene is classified as upregulated in a disease if the average concentration of all samples of that disease is significantly higher (adj P-value<0.05) than the average concentration of samples of all diseases as measured by PEA . For gender specific diseases the analysis includes only samples corresponding to the same gender from the other diseases.

The disease(s) the gene is associated with and able to predict according to glmnet prediction models. To be included the gene has to be upregulated according to differential expression analysis and have more than 50% overall importance as indicated by the prediction models.

All genes with at least one predicted secreted isoform have been annotated and classified with the aim to determine if the corresponding protein(s) are:

• secreted into blood
• locally secreted
• or actually being attached to membrane or retained in intracellular locations like mitochondria, endoplasmatic reticulum (ER), Golgi apparatus or lysosomes.

Detection or not of the gene in blood, based on spectral count estimations from a publicly available mass spectrometry-based plasma proteomics data set obtained from the PeptideAtlas.

Detection or not of the gene in blood, based on proximity extension assays (Olink) for a longitudinal wellness study covering 76 individuals with three visits during two years.