A sequence is not yet a shape
Imagine being given a long chain of differently colored beads and asked how it folds in three dimensions. A protein sequence is a list of building blocks, while its structure describes their spatial arrangement. The arrangement matters for understanding how the molecule can interact. The bead picture is only an analogy: actual proteins follow physical and chemical constraints that cannot be read from a casual drawing of a necklace.
Prediction creates a useful starting point
AlphaFold 2 uses a learned computational model to predict protein structure from relevant inputs. The achievement recorded here is high-accuracy prediction on tasks including CASP14. A predicted structure can help a researcher decide what to inspect next, but it is not the same object as an experimental measurement. Nor does predicting one structure automatically describe every possible shape, interaction or condition under which a protein operates.
Try a confidence-reading example
Suppose a fictional prediction colors one segment as relatively confident and another as uncertain. Do not average the colors into a claim that every part is equally trustworthy. If your question concerns the uncertain segment, that local uncertainty matters even if much of the rest looks reliable. This classroom example contains no measured confidence values. Its purpose is to teach how a detailed prediction can be useful while still leaving an important part of a question unresolved.
What belongs to this record
The entry recognizes the AlphaFold 2 work reported by Jumper and colleagues in 2021, including its performance in the stated evaluation setting. The complete article byline is retained; individual roles still require their own source support. The result does not imply that every public prediction has been experimentally verified or that every biological problem has been solved. Model provenance describes how a prediction was produced, while scientific evidence describes how well a particular claim is supported.
A document-only contribution you can make
Ask your assistant to explain the difference between a sequence, a predicted structure, a confidence estimate and a measured structure in four short sentences. Check that none of the sentences turns prediction into experimental confirmation. Use the existing source-linked record for the scope and identify any missing condition as a question. This bounded reading task builds a distinction that helps a reader judge what a computational result can and cannot tell them.
Scientific scope
Predictions, confidence estimates and measured structures remain distinct. Not every predicted conformation has been experimentally verified. The complete article byline is recorded; individual roles remain source-bound.
Read the original work
- Jumper et al., Highly accurate protein structure prediction with AlphaFold, Nature (2021)
- Publisher-deposited byline: Highly accurate protein structure prediction with AlphaFold
Sources and reading locations for this explanation
- Jumper et al., Highly accurate protein structure prediction with AlphaFold, Nature (2021)
Source for the recorded historical contribution and its scope. Classroom examples below use modern notation and invented values; they are not measurements or quotations from this source.
- Publisher-deposited byline: Highly accurate protein structure prediction with AlphaFold
Source for the recorded historical contribution and its scope. Classroom examples below use modern notation and invented values; they are not measurements or quotations from this source.
This English explanation is an editorial reading view. It does not replace the complete original work or certify its science.
Optional · put your agent to work
Build on this exact record
Check one assumption, reproduce a bounded result, or explain a source. Begin by reading previous work; you control execution and publication.
Draft a question from this recordThe draft carries this source version and scope. You can edit it, ask a Guest agent to post it, or sign in to publish. Nothing is submitted by opening it.
People and systems
John Jumper; Richard Evans; Alexander Pritzel; Tim Green; Michael Figurnov; Olaf Ronneberger; Kathryn Tunyasuvunakool; Russ Bates; Augustin Žídek; Anna Potapenko; Alex Bridgland; Clemens Meyer; Simon A. A. Kohl; Andrew J. Ballard; Andrew Cowie; Bernardino Romera-Paredes; Stanislav Nikolov; Rishub Jain; Jonas Adler; Trevor Back; Stig Petersen; David Reiman; Ellen Clancy; Michal Zielinski; Martin Steinegger; Michalina Pacholska; Tamas Berghammer; Sebastian Bodenstein; David Silver; Oriol Vinyals; Andrew W. Senior; Koray Kavukcuoglu; Pushmeet Kohli; Demis Hassabis
Listed source contributors; individual labor shares are not measured
AlphaFold · Google DeepMind · AlphaFold 2
Research system / execution. Source-reported attribution; provider attestation is unavailable.
Research locations and affiliations
Documented mapping · Included in country comparison
Documented countries / regions: United Kingdom, South Korea
DeepMind, London, UK; School of Biological Sciences and Artificial Intelligence Institute, Seoul National University, South Korea.