Statistics
Data distribution across ADMETatlas
A visual summary of peptide records, endpoint evidence, context signals, product context, and reference coverage in the current public release.
Stack
Statistics
A visual summary of peptide records, endpoint evidence, context signals, product context, and reference coverage in the current public release.
The statistics page is a distribution view. It shows what the current atlas contains and how records are arranged across scientific modules, evidence roles, and product context. Each chart states its denominator as n = …, and any percentages are shares of that denominator in the current release.
What kind of peptides the atlas holds: sequence length, residue makeup, and chemical topology. These distributions describe the molecules themselves, before any endpoint evidence is layered on.
Residue counts for canonical standard amino-acid sequences. Cyclic peptides and other modified forms are summarized separately, below.
Residues tallied across all standard amino-acid sequences, colored by side-chain class.
Topology class summarizes sequence/form representation, not developability quality.
Monomer count for cyclic peptides, using the source's authoritative monomer length. A further 5,162 modified or text-encoded forms from other sources use heterogeneous residue notations and are not length-binned.
Endpoint evidence is grouped by the five formal ADMET modules used across the site. The sunburst shows module-to-endpoint hierarchy, while the ranking shows where evidence volume is concentrated.
Outer slices are endpoints; inner slices are the five formal modules.
Top 12 of 17 formal endpoints by record count. Labels link to endpoint pages below.
How the actual measured values are distributed for the highest-volume numeric endpoints. Each histogram is binned within a single coherent unit; endpoints whose values mix incomparable units are deliberately not shown here.
Reported values in log permeability (source-reported). Shaded band marks High permeability ≥ −6.0.
Reported values in % hemolysis.
Reported values in % cytotoxicity.
Reported values in µM (log scale).
Reported values in µM (log scale).
The experimental conditions under which the endpoint evidence was measured, normalized to controlled scientific vocabularies. Each measurement is comparable only within a matching organism, biological matrix, and route — these distributions show the assay landscape behind the numbers. Counts cover measurements where the source recorded the field; bars omit measurements with no recorded value.
Organism (or organism-derived system) in which each endpoint measurement was made.
The biological material the peptide was measured in (plasma, cell culture, tissue, buffer, …). Values are only comparable within the same matrix.
Administration route for in-vivo / dosing studies. 'Multiple routes' marks measurements that report more than one route.
How the marketed and tracked peptide products in the atlas are described, kept separate from endpoint evidence.
Routes are release-level product descriptors and should not be interpreted as endpoint evidence.
Priority bands help orient flagship examples versus broader catalog records.
How traceable the evidence is by source type, and how broadly each endpoint is independently corroborated across distinct external datasets. Individual upstream database names are kept out of the charts and listed in references and downloads instead.
Each reference carries one scientific source class. ADMETatlas standardizes curated datasets and directly-extracted primary literature into one comparable evidence model, so every value keeps a traceable origin.
Bands group endpoints by how many distinct external datasets report concordant evidence for them — more datasets mean broader independent corroboration, not a quality score. (The same primary study can reach the atlas through more than one dataset.)
Curation mechanics behind the numbers: how measurements are classed by scientific role and value form, and how context signals attach. These support traceability and are collapsed by default to keep the page science-forward.
Stacked bars show how each of the top 12 of 17 endpoints is composed.
Numeric values, ranges, categorical values, and text observations remain separate.
Context signals support interpretation around trials, product identity, and reviews; they are not formal endpoint measurements.
Signals attach at the product level, to the linked peptide identity, or — when identity is not yet resolved — to the source record alone.
