The Human Protein Atlas-Spatial localization of the human proteome in health and disease

Abstract

For a complete understanding of a system's processes and each protein's role in health and disease, it is essential to study protein expression with a spatial resolution, as the exact location of proteins at tissue, cellular, or subcellular levels is tightly linked to protein function. The Human Protein Atlas (HPA) project is a large-scale initiative aiming at mapping the entire human proteome using antibody-based proteomics and integration of various other omics technologies. The publicly available knowledge resource www.proteinatlas.org is one of the world's most visited biological databases and has been extensively updated during the last few years. The current version is divided into six main sections, each focusing on particular aspects of the human proteome: (a) the Tissue Atlas showing the distribution of proteins across all major tissues and organs in the human body; (b) the Cell Atlas showing the subcellular localization of proteins in single cells; (c) the Pathology Atlas showing the impact of protein levels on survival of patients with cancer; (d) the Blood Atlas showing the expression profiles of blood cells and actively secreted proteins; (e) the Brain Atlas showing the distribution of proteins in human, mouse, and pig brain; and (f) the Metabolic Atlas showing the involvement of proteins in human metabolism. The HPA constitutes an important resource for further understanding of human biology, and the publicly available datasets hold much promise for integration with other emerging efforts focusing on single cell analyses, both at transcriptomic and proteomic level.

Keywords: Human Protein Atlas; antibodies; cells; proteomics; single-cell; tissues; transcriptomics.

FIGURE 1

The global structure of the Human Protein Atlas web portal. Exploration of the Human Protein Atlas data is guided through a systemic and a gene‐centric approach, both subdivided into six interconnected sub‐atlases. The systemic approach entails exploration of various sub‐atlas proteomes, each specific to a group of genes belonging to a certain location or activity within the body, such as an organ, blood, or genes involved in metabolism. The sub‐atlas pages are accessed by clicking on either of the sub‐atlas images on the startpage. The other option is to search for specific genes using the startpage search field that can be combined with various different filtering options. The search leads the visitor to a search result page, wherein the gene of interest can be selected to access gene‐specific data, both summarized in the gene summary page and presented in depth in the different gene‐specific sub‐atlases. In addition, the database contains a page with data concerning SARS‐CoV‐2 relevant proteins, as well as downloadable datasets and educational material such as dictionaries, found in the menu (blue frames)

FIGURE 2

The Tissue Atlas—integration of mRNA expression with IHC. (a) Overview of the 37 main organ systems with RNA expression levels based on three different datasets (HPA, GTEx, and FANTOM5), used for classification of tissue specificity and tissue distribution. (b) Pie charts showing the number of genes belonging to each of the categories for tissue specificity and tissue distribution. (c) IHC staining of zona pellucida binding protein (ZPBP) in testis tissue, outlining the eight different cell types present in the sample: 1: spermatogonia; 2: preleptoteine spermatocytes; 3: pachytene spermatocytes, 4: round/early spermatids; 5: elongated/late spermatids; 6: Sertoli cells; 7: Leydig cells; 8: peritubular cells. ZPBP is highly expressed in round/early spermatids and elongated/late spermatids

FIGURE 3

The Pathology Atlas—association of gene expression with clinical outcome. (a) Interactive survival scatter plot of cyclin dependent kinase 1 (CDK1) in lung cancer. The plot shows the clinical status (i.e., dead or alive) for all individuals in the patient cohort. The top kernel density plot demonstrates the expression level (FPKM) distribution among dead (red) and alive (blue) patients. The right density plot shows the data density of the survived years of dead patients with high and low expression levels, respectively, stratified using the cutoff indicated by the vertical dashed line. The cutoff is automatically generated, representing the lowest P score. (b) The P score landscape (black curve; left axis) is shown together with dead median separation (red curve; right axis), highlighting the difference in median mRNA expression between patients who have died with high and low expression, respectively

FIGURE 4

The Brain Atlas—exploration of gene expression in the mammalian brain. (a) Overview of the 10 main regions with RNA expression data from all three species (human, pig, and mouse), used for classification of genes according to expression levels. (b) Pie charts showing the number of genes belonging to each of the categories for tissue specificity and tissue distribution, comparing brain to other human tissues. (c) IHC images of ANK1 and TFAP2B that on a body‐wide level are elevated in other tissue types, but in the brain showed regional expression only in cerebellum. (d) Immunofluorescent images of whole mouse brain, as well as five specific regions detailing the protein expression of NECAB1 at a cell type‐specific level

FIGURE 5

The Blood Atlas—exploration of gene expression in human blood cells. (a) A schematic visualization of the tree of hematopoiesis is found in the Blood Atlas, including six lineages and 18 blood cell types that are clickable to enable access to expression profiles with information regarding expression‐related specificity and distribution. (b) RNA expression levels of known protein markers for specific blood cell types, and (c) representative IHC images of corresponding protein expression patterns in human tissues

FIGURE 6

The Human Secretome—annotation of actively secreted proteins. (a) The complete set of predicted actively secreted proteins were annotated and divided into nine different classes based on their final location in the body. Each class was analyzed regarding (b) protein function, (c) expression specificity and distribution, and (d) tissue of origin among tissue enriched proteins (exemplified by the class “secreted to blood”). (e) Spatial IHC analysis of F2 in five tissues, a protein that is annotated as secreted to blood by the liver, with exclusive mRNA expression in the liver

FIGURE 7

The Metabolic Atlas—integration of metabolic pathways with tissue‐specific proteomics. Features of the Metabolic Atlas (metabolicatlas.org) have been integrated into HPA as (a) a sub‐atlas with summarized information and maps regarding all human metabolic pathways (example showing the mitochondrial subcellular compartment with the Urea cycle indicated by a brown box) and (b) integrated into the Tissue Atlas part of the exploration of individual genes, including a summary of gene‐associated metabolic information, maps of associated metabolic pathways and heatmaps showing the mRNA expression of all pathway‐associated genes in 37 tissues