Biophysical Proteome Atlas

How to use the atlas

A short tour from the front page to a downloaded prediction, using the human tumour suppressor P53 (P04637) as the worked example. It covers how to navigate the site — for what the predictions mean and how they are computed, see the methodology and the publications it cites.

1. Find a protein

If you already know your protein, the search box on the home page is the fastest way in: give it a UniProt accession (P04637) or an entry name (P53_HUMAN) and it looks across every published proteome. Searching by gene or protein name instead works from each proteome's entry list — step 4 below.

The atlas home page: the search box above five counters summarising domains, kingdoms, proteomes, entries and views.
The home page. The search box accepts a UniProt accession or entry name; the counters below say what the atlas currently holds.

The navigation bar is the same on every page. Organisms opens a menu of every published proteome — the one you are currently reading is underlined in it — and while you are inside a proteome an Entries item appears beside it, leading to its entry list.

2. Pick an organism

Browsing instead? Below the search, the home page lists every published proteome. Each card states how many entries it holds and which UniProt release it was imported from; Open proteome leads to the proteome's own page, Entry list straight to its table of entries. Every proteome carries the same predictions, so results are comparable between species.

The organism picker: one card per published proteome, each with its entry count, UniProt release and links.
The organism picker on the home page.

3. Explore a proteome

A proteome's home page carries a search box scoped to that organism — this one also matches gene and protein names, and suggests entries as you type. Next to it: browse the full entry list, or jump to one chromosome where the proteome has them. Further down, the page describes the proteome itself — what UniProt says about it, and exactly what this atlas includes of it (the reviewed Swiss‑Prot entries, with the reviewed/unreviewed split drawn out).

The human proteome page: hero, per-proteome search card with entry, gene and chromosome counts.
The human proteome's home page. The search covers this proteome only and also matches gene and protein names.

4. Browse the entry list

The entry list is the proteome as a table: one row per protein. Filter it by accession, protein or gene name — and by chromosome, where the organism has them. Click a column heading to sort, and click it again to reverse the order; on a small screen the rows become cards and a Sort by control appears in the filter bar instead. The Actions column offers each row's files under Downloads and the same entry as machine-readable JSON under REST API. Filtering the human list by the gene name TP53 finds our worked example; its accession opens the entry page.

The human entry list filtered to TP53: a filter bar, then a sortable table of accessions with an Actions column of Downloads and REST API controls.
The human entry list filtered by gene name — TP53 finds the worked example. Every column heading with arrows sorts.

5. Read an entry

The entry page is where the predictions live. At the top: the accession and entry name, and links out to the same protein on PDBe‑KB, UniProtKB, AlphaFoldDB, InterPro and Scop3P. Below, the content is organised in tabs:

  • Overview — every predictor on one plot, per residue. Click a name in the legend to toggle that line; drag to zoom. For P53 the plot shows the shape the protein is known for — a rigid, folded core with flexible, disordered stretches at both ends.
  • Interpretation — each residue classified as order, transition or disorder.
  • Values and Statistics — the numbers behind the plots, per residue and summarised.
  • Sequence — the FASTA record the predictions were computed from.
  • Plots — the same predictions, one plot per predictor.
  • PSP — phase-separation propensity.
  • Visualization 1D‑3D — the predictions painted onto a 3D structure, where one is available.
  • Help — the in-page legend: what each line is and the thresholds that matter when reading it.
The P04637 entry page: external resource links, the tab bar, and the overview plot of every predictor across the sequence.
The P04637 entry. The overview plot carries every predictor; the tabs above it break the same data out.

6. The first view computes the predictions

Outside the human proteome, most entries are imported as sequences only and their predictions are computed the first time somebody opens them — usually well under a minute. The page says so, shows the progress, and reloads itself when the results are ready; nothing needs to be requested or configured. Every later visit is served the stored result directly.

An entry before its first view: the page explains that predictions are generated on first view, with the sequence already shown.
An entry nobody has opened yet. The sequence and provenance are already complete; the predictions follow.

7. Download the data

Each entry page carries a Downloads panel: the sequence as FASTA, the biophysical predictions as JSON (per-residue values for every predictor), the disorder interpretation as JSON, and a lookup of experimental and predicted 3D structures. The same files are available without opening the entry, from the Downloads menu in the entry list's Actions column.

The entry page's Metadata and Downloads panels: sequence, biophysical predictions, disorder interpretation and 3D structures.
The Downloads panel, next to the entry's provenance metadata.

8. Use it from code

Everything above is also served as JSON. The quickest way in is the REST API link on any entry list row; the whole interface — entry metadata, predictions, per-organism entry lists — is documented on the API page, machine-readable schema included. The worked example, from a terminal: 

curl https://bio2byte.be/proteome/api/protein/P04637/

Where to go deeper

This tour stops where the science starts. The methodology page explains each predictor and cites the publications behind it, the entry page's own Help tab explains how to read the plots, and the status page shows what the atlas is storing, serving and computing right now.