Bioinformatics

Computational organization and analysis of biological information. Comparing sequences, annotating genes and analyzing expression measurements requires both algorithms and biological assumptions; a numerical similarity score alone does not establish a shared function.

RELION is software for statistically regularized reconstruction and classification of macromolecular structures from single-particle cryo-electron microscopy images. A particle image is a noisy projection with initially uncertain orientation and displacement. The reconstruction must infer those hidden quantities while estimating the three-dimensional structure. ^[ref]

The Bayesian framework combines an image likelihood with prior information that limits unsupported high-resolution detail. Iterative classification can separate subsets consistent with different structures rather than averaging every particle into one object. Regularization does not guarantee a correct result: starting models, image quality, heterogeneity and validation still matter. A sharp-looking reconstruction is a model inferred from measurements, not a direct three-dimensional photograph.

A plain-text representation of biological sequences. A record begins with a greater-than sign and an identifying header, followed by sequence characters; the format stores the sequence but does not by itself certify its origin, accuracy or biological interpretation.

An interface displaying annotations and measurements at positions along a reference genome. Tracks align features such as genes and experimental signals in a common coordinate system; the reference assembly and coordinate convention are essential to interpreting overlap.

Computational analysis of biological three-dimensional structures. It connects structural comparison, prediction and annotation with questions about molecular relationships or function, while retaining the distinction between a predicted model and experimental evidence.

Inference of a protein’s three-dimensional structure from information such as sequence, related sequences and known structures. A useful prediction includes an assessment of uncertainty; a confident local fold does not automatically establish a molecular interaction or biological function.

An AlphaFold confidence score for local structural accuracy, usually reported on a 0–100 scale. It is not a probability that a complete structure is correct.

A family of computational systems for predicting biomolecular structures. Its second major version showed that learned sequence and structural relationships can support highly accurate protein models, while confidence and biological context remain necessary for interpretation.^[ref]

The protein-structure prediction system evaluated in CASP14 and reported in 2021. It combines information from related sequences with learned structural relationships, producing coordinates together with confidence estimates rather than a direct experimental structure.^[ref]

Expected positional error at one residue after predicted and true structures are aligned on another residue. Low within-domain values with high between-domain values support more confident internal geometry than relative domain placement. PAE is not a physical residue–residue distance and need not be symmetric.^[ref]