Predicted Aligned Error¶

An asymmetric residue-pair matrix estimating the expected positional error at one residue when a predicted protein structure is aligned on another residue's local frame.

Version: v2 · 2026-09-06 · History
Domain-specific #: 2521
Origin domain: structural bioinformatics
Subdomain: AlphaFold confidence outputs
Aliases: PAE, Predicted alignment error

Core Idea¶

Predicted Aligned Error is an asymmetric residue-pair matrix estimating the expected positional error at one residue when a predicted protein structure is aligned on another residue's local frame. ^[1]

For residues x and y, PAE(x,y) estimates the expected positional error at residue x, measured in angstroms, if predicted and true structures were aligned on residue y's local frame. The matrix is directional: aligning on y and measuring x is not the same conditional operation as aligning on x and measuring y. Low inter-domain blocks support confidence in relative placement; high blocks warn that apparently precise packing should not be interpreted.

The operative boundary is exact: The alignment-conditioned pairwise confidence matrix for relative protein-domain placement remains uncovered. The abstraction is therefore not the topic named by its field, but the reusable role structure specified below.

Structural Signature¶

Sig role-phrases:

the predicted structure — the coordinate model being assessed
the hypothetical true structure — the unavailable reference used in the training target
the aligned-on residue y — the local frame fixing superposition
the measured residue x — the position whose post-alignment error is estimated
the expected distance error — the model-weighted mean over aligned-error bins
the ordered residue pair — the source of matrix directionality and possible asymmetry
the PAE matrix — all ordered pair estimates displayed as a heatmap
the domain block pattern — within- and between-domain confidence structure
the model-version calibration — the learned confidence output's dependence on system and training

Recognition test. A case qualifies only when its roles can be mapped to the declared the predicted structure, the hypothetical true structure, the aligned-on residue y, the measured residue x, and when the characteristic boundary conditions are preserved. Surface vocabulary or a loose analogy is insufficient.

What It Is Not¶

Not an observed error against an experiment. PAE is predicted confidence when a true structure is usually unavailable.
Not pLDDT. pLDDT is principally per-residue local confidence, not alignment-conditioned pairwise placement.
Not an ordinary symmetric distance matrix. the ordered alignment condition permits asymmetry.
Not proof that high-PAE domains do not interact. High PAE means relative placement is uncertain, not necessarily biologically absent.
Not a substitute for experimental validation. Confidence guides interpretation but does not provide ground truth.
Not one scalar model score. The matrix preserves pairwise structure lost in a single average.

Scope of Application¶

The abstraction has a bounded but recurring habitat. These are literal applications of the same domain machinery, not cross-domain metaphors. ^[2]

Domain packing. between-domain blocks show confidence in relative orientation and position.
Flexible linkers. high off-diagonal error can reveal uncertain relative motion between locally confident domains.
Protein complexes. inter-chain blocks provide one view of relative subunit placement, alongside interface-specific metrics.
Model triage. regions suitable for local interpretation are separated from uncertain global topology.
Construct design. domain boundaries and flexible segments can inform experimental planning cautiously.
Confidence visualization. heatmaps expose structure that per-residue coloring cannot.

Clarity¶

Always declare the axis convention. In the AlphaFold Database presentation, color at (x,y) is expected error at x after alignment on y. The diagonal is low by construction and is not evidence that the whole model is accurate. Interpretation belongs mainly to off-diagonal block structure.

A useful audit proceeds in order: identify the candidate roles, verify their types and quantifiers, apply the recognition test, and then test every stated exclusion. If a case supplies only the broad parent pattern while dropping the domain accent, it is not Predicted Aligned Error.

Manages Complexity¶

PAE reduces a family of alignment-conditioned uncertainties to one matrix without collapsing them to a scalar. It separates local fold confidence from global assembly confidence and makes directionality, domains, chains, and flexible connections visible in one diagnostic surface.

The compression remains accountable because every simplification has a named validity condition. A user can ask which role is missing, which assumption fails, and which neighboring abstraction should replace the candidate instead of treating the label as an unanalyzed bundle.

Abstract Reasoning¶

R1. Record which axis is aligned and which is measured.

R2. Inspect off-diagonal blocks rather than relying on the dark diagonal.

R3. Compare PAE with pLDDT and interface metrics because they answer different questions.

R4. Treat high PAE as uncertainty, not a negative biological claim.

R5. Preserve model version and confidence calibration when comparing outputs.

The reasoning pattern is deliberately typed: definitions establish identity, calculations or constructions establish consequences, and empirical or institutional evidence establishes whether a real case instantiates the roles. One kind of support cannot silently substitute for another.

Knowledge Transfer¶

PAE transfers literally among structure-prediction systems that define the same alignment-conditioned expected error. Generic pairwise uncertainty is the parent idea. Using the acronym for any symmetric contact score or coordinate RMSD drops the defining local-frame conditioning.

The transfer boundary follows from the classification test: The output recurs across AlphaFold structure assessment, while residue-axis convention, frame alignment, expected-distance bins, asymmetry, domain confidence, and model-version calibration remain constitutive. The safe portable move is to name the broader parent when the home-domain machinery is absent and to retain the domain name only when literal recognition succeeds.

Examples¶

Canonical: two-domain protein¶

A predicted protein has low PAE within residues 1–120 and within residues 150–260, but high PAE between the two blocks. Each domain may be locally credible, while their relative orientation is uncertain. Rendering the one coordinate model as a rigid two-domain assembly would overstate what the prediction supports. ^[1]

Mapped back: the PAE matrix; the domain block pattern; the expected distance error; the ordered residue pair.

Applied / In Practice: axis-aware inspection¶

A researcher selects cell (x,y) in a PAE plot and states: after aligning prediction and truth on residue y's frame, the model expects residue x to have the displayed positional error. They then inspect the transposed cell separately. Differences are not a plotting defect; they reflect two different alignments and measured positions. ^[2]

Mapped back: the aligned-on residue y; the measured residue x; the ordered residue pair; the model-version calibration.

Structural Tensions¶

T1: Local confidence versus global topology. Domains can be individually accurate while their packing is unconstrained. Diagnostic: Are local and inter-domain blocks being interpreted separately?

T2: Matrix richness versus visual simplification. Heatmaps reveal patterns but palettes and clipping can hide numerical differences. Diagnostic: Have raw values and scale limits been checked?

T3: Directionality versus symmetric intuition. Pairwise displays look like distance matrices even though alignment makes entries ordered. Diagnostic: Which residue defines the alignment frame?

T4: Predicted uncertainty versus empirical error. PAE can be calibrated on benchmarks but remains a model output for any new target. Diagnostic: Is the statement about expected confidence or measured accuracy?

T5: High uncertainty versus biological flexibility. Flexible linkers can produce high PAE, but model uncertainty and real conformational heterogeneity are not identical. Diagnostic: What independent evidence supports flexibility?

T6: Domain autonomy vs prime reduction. Prediction Error and uncertainty are broad, but local-frame alignment, ordered residue pairs, and protein-domain heatmaps define PAE. Diagnostic: Would a generic residual or covariance matrix preserve the same semantics? If not, retain the domain node.

Structural–Framed Character¶

The five-criterion aggregate is 0.30 (mixed-structural). The classification is reasoned rather than cosmetic:

Vocabulary travels — mixed (0.50). The operative vocabulary retains the home-domain types named in the Structural Signature even when a thinner parent pattern travels.
Evaluative weight — structural (0.00). The score records whether applying the abstraction requires a normative or interpretive judgment in addition to structural recognition.
Institutional origin — mixed (0.50). The score records whether the abstraction is constituted by a scholarly, legal, technical, or administrative convention rather than merely discovered in nature.
Human-practice bound — structural (0.25). The score records how far the named roles depend on a human practice, measurement regime, language, or institution.
Import versus recognize — structural (0.25). Beyond its home habitat, use of the name increasingly becomes import by analogy rather than recognition of the same mechanism.

The portable skeleton is: condition an uncertainty estimate on a reference alignment and retain ordered pairwise errors so local and global confidence separate. That skeleton belongs to the related parent abstractions; it does not make the fully accented node a prime. Its character: mixed-structural, with a real structural core whose recognition remains bounded by domain-specific types and validity conditions.

Structural Core vs. Domain Accent¶

This section decides why Predicted Aligned Error is a domain-specific abstraction rather than a prime.

Structural core: Condition an uncertainty estimate on a reference alignment and retain ordered pairwise errors so local and global confidence separate. This relational skeleton can recur outside the home domain and is the part legitimately carried by broader primes.

Domain accent: Protein residues, predicted and true local frames, angstrom error bins, alphafold confidence heads, domain packing, chains, and heatmaps. Remove those types and constraints and the result may still resemble the skeleton, but it is no longer recognized as this named abstraction.

Why it does not clear the prime bar: Conditional uncertainty travels broadly; PAE is the exact structure-prediction output whose axes and alignment frame determine interpretation. Cross-domain transfer is therefore routed through the parents, while the named entry remains available for precise in-domain diagnosis.

Prediction Error. is the broad discrepancy concept.
Calibration. governs whether predicted errors match benchmark frequencies.
Uncertainty. is the parent quantity retained in pairwise form.

These are prose relations only. They do not create structured DAG edges, and placement must still pass the live endpoint, redundancy, and cycle checks recorded in the bundle's placement memo.

Relationships to Other Abstractions¶

Current abstraction Predicted Aligned Error Domain-specific

Parents (1) — more general patterns this builds on

Predicted Aligned Error is a kind of Uncertainty Prime

Uncertainty. is the parent quantity retained in pairwise form.

Hierarchy path (1) — routes to 1 parentless root

Predicted Aligned Error → Uncertainty

Neighborhood in Abstraction Space¶

Predicted Aligned Error sits in a sparse region of the domain-specific corpus (73^rd percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (1565 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-09-08

Not to Be Confused With¶

pLDDT. per-residue local confidence. Tell: Is the question local fold quality or relative placement?
RMSD. observed coordinate discrepancy after a chosen superposition. Tell: Is a true reference structure available?
Contact probability. probability that residues are spatially near. Tell: Is the value proximity or alignment-conditioned error?
Distance matrix. pairwise distances within one structure. Tell: Does each entry depend on aligning a local frame?
ipTM. a scalar confidence score for interfaces or complexes. Tell: Is pairwise block detail retained?

References¶

[1] AlphaFold Protein Structure Database, “Frequently Asked Questions”, EMBL–EBI and Google DeepMind. registry ↩a ↩b

[2] EMBL–EBI Training, “PAE: A Measure of Global Confidence in AlphaFold2 Predictions”. registry ↩a ↩b