Protein structure guides

Guides on predicting, scoring, and comparing protein structures with AlphaFold, Boltz, and related tools.

Protein ribbon beside separate license documents labeled Code and Weights.

Structures ·

AlphaFold license and commercial use

AlphaFold 2 and the AlphaFold Database allow commercial use, while the AlphaFold 3 weights and AlphaFold Server do not. See what each license allows and which models, including Boltz-2, Chai-1, RFdiffusion and ESM, companies can use.

Model and reference protein complexes overlaid for comparison of their interface.

Structures ·

DockQ score explained

DockQ scores how closely a predicted protein complex matches a reference interface, from 0 to 1, with CAPRI cutoffs at 0.23, 0.49 and 0.80. Learn how it combines fnat, iRMSD and LRMSD, and how pDockQ estimates it with no reference.

Protein fold with a magnified view of overlapping nonbonded atomic envelopes.

Structures ·

What the MolProbity score and clashscore mean

The MolProbity score combines clashscore, rotamer outliers and Ramachandran favored residues into one number on the scale of crystallographic resolution. Learn what counts as good and how AlphaFold models score.

Two folded protein domains joined by a linker, with ghost outlines showing uncertain relative placement.

Structures ·

What is PAE, and how to read a PAE plot

PAE (predicted aligned error) is the pairwise confidence matrix that AlphaFold, Boltz and Chai-1 write next to a predicted structure. Learn how to read its blocks and stripes and use it for domains and interfaces.

Schematic protein backbone with phi and psi rotations around the N–Cα and Cα–C bonds.

Structures ·

How to read a Ramachandran plot

A Ramachandran plot shows the phi and psi backbone angles of every residue in a protein. Learn what the regions mean, what percentages a good structure has, and how to tell a model error from a strained residue.

Superimposed protein folds with dotted connectors showing differences at corresponding positions.

Structures ·

What is RMSD, and what counts as a good value

RMSD is the average distance between matched atoms of two structures, in ångströms. Learn the formula, when to superimpose, and what values count as good for proteins, docked poses and molecular dynamics.

Protein model with callouts for local confidence, relative domain placement and atomic geometry.

Structures ·

How to tell if a predicted protein structure is reliable

A predicted structure is reliable for a question when the scores that answer that question agree. Read pLDDT, PAE, pTM and ipTM in order, rescore interfaces with ipSAE, then check accuracy and geometry with DockQ and MolProbity.

Two protein chains with low-PAE residue pairs highlighted at their interface and faint flexible tails

Structures ·

ipSAE explained

ipSAE is an interface confidence score for AlphaFold2, AlphaFold3 and Boltz that ignores disordered tails. Learn how it is calculated, why it beats ipTM and what 0.6 to 0.8 mean.

Two-chain protein complex with pTM marking the overall structure and ipTM marking relationships between chains

Structures ·

ipTM vs pTM: judging predicted complexes

pTM scores the whole predicted structure, while ipTM scores only how chains sit relative to each other. Learn what the 0.5, 0.6 and 0.8 cutoffs mean and why disordered tails lower ipTM.

Protein ribbon with a local geometry callout and an uncertain tail, illustrating per-residue confidence

Structures ·

What is pLDDT, and what counts as a good score

pLDDT is the per-residue confidence score that AlphaFold, ESMFold and Boltz attach to predicted structures. Learn what the 90, 70 and 50 cutoffs mean and why a low score is not always a failure.

Superimposed protein folds with a shared core and displaced loop, illustrating TM-score and RMSD

Structures ·

TM-score vs RMSD

RMSD and TM-score both compare two protein structures, but they answer different questions. Learn how each is calculated, what values count as good, and which one to use for prediction, fold search, docking or design.

Engraved protein folds illustrating a collection of predicted structures.

Structures ·

AlphaFold database statistics [2026]

AlphaFold DB lists 261,552,403 predicted models as of September 24, 2026. See the collection breakdown, human proteome coverage, and how the database compares with the PDB.

Conceptual diagram of semaglutide acting at the GLP-1 receptor and tirzepatide acting at GLP-1 and GIP receptors; molecular shapes are illustrative.

Drugs ·

Semaglutide vs tirzepatide

Semaglutide (Ozempic, Wegovy) vs tirzepatide (Mounjaro, Zepbound): head-to-head weight loss and diabetes results, side effects and U.S. prices, with sources.

Conceptual scientific illustration of a peptide with a branched side chain associating with albumin.

Drugs ·

What is semaglutide?

Semaglutide’s amino-acid sequence, molecular weight, GLP-1 mechanism and oral delivery, with our molecular calculations and audit of two PDB structures.

Ink illustration of archived protein structure records and a crystal representing the Protein Data Bank.

Structures ·

Protein Data Bank statistics 2026: size and coverage

Current Protein Data Bank statistics for archive size, protein sequence coverage, experimental methods, growth, usage, computed models, and structure data formats.