Skip to content
PaperFren

Can an SVM predict which amino acids touch in a folded protein?

Open paper intelligence

A support vector machine using many sequence-derived features predicted residue contacts better than the previous leading method, though accuracy remained low overall.

Source

Improved residue contact prediction using support vector machines and a large feature set

Cheng J, Baldi P · BMC bioinformatics · 2007

doi.org/10.1186/1471-2105-8-113Read the full paper ↗133 citationscc by

Study at a glance

Design
Computational / modelling — RBF-kernel SVM classifies residue pairs as in contact or not; trained and tested on the same split as CMAPpro, then assessed on CASP7 de novo domains.
N
N=48 · 48 test proteins (485 training proteins); the CASP7 comparison used 13 released de novo domains.
Population
Non-redundant protein chains (under 25% pairwise sequence identity) and CASP7 targets
Outcome
Accuracy (specificity) and coverage (sensitivity) of predicted medium- and long-range residue contacts

Structured fields used in claim comparison tables when every cited study has a complete layer.

What they did

The authors trained SVMcon, a support vector machine with a Gaussian (RBF) kernel, to decide whether two residues at least six positions apart in a protein sequence lie within 8 Å of each other. Each residue pair was described by hundreds of features: sequence profiles in windows around both residues, predicted secondary structure and solvent accessibility, pairwise mutual information, contact potentials and whole-protein composition. They trained on 485 proteins, tested on 48 against CMAPpro, and entered the blind CASP7 competition.

What they found

At the point where sensitivity equals specificity SVMcon reached 27.1%, about four points higher than CMAPpro and roughly nine times a random guess. Contacts in proteins with beta-sheets were predicted more accurately than those in all-alpha proteins. In CASP7 it had 27.7% accuracy at a sequence separation of at least 12, second among the eight predictors compared.

The limits

What it doesn't show

Even the best predictions are mostly wrong at longer ranges, so this is a step toward, not a solution for, building 3D structures from contacts. The CASP7 set was only 13 domains, and the authors warn against over-interpreting the ranking and note their evaluation differs slightly from the official one. Training took days, so only a crude feature-removal check was possible, and the paper does not test whether the predicted contacts actually improve 3D modelling.

Key terms

Residue contact
Two amino acids whose alpha-carbons are within 8 Å of each other in the folded protein.
Support vector machine
A classifier that finds the boundary with the widest margin between classes, possibly after mapping inputs to a higher-dimensional space with a kernel.
RBF kernel
A similarity function that decays with the squared distance between two inputs, letting an SVM draw curved decision boundaries.
Break-even point
The decision threshold where sensitivity equals specificity, used to compare predictors on equal footing.
CASP
A blind community experiment where methods predict protein structures before the true structures are released.

Flashcards

1 / 9

0 of 9 answers reviewed

Research intelligence for this paper

See its role on concept claims, tensions it is part of, placement history, and related discoveries.

Open paper intelligence

Quiz yourself

1 / 5

What kind of model is SVMcon?

Common questions

Why only predict contacts six or more residues apart?

Short-range contacts are largely explained by local secondary structure; the informative, hard cases are medium- and long-range pairs.

Why compare with a coin-flip baseline?

Contacts are rare, so a random guesser tuned to the base rate shows how much real signal the model adds; SVMcon was about nine times better.

Why did the authors throw away most negative examples?

Non-contacts vastly outnumber contacts, so they kept a random 5% of negatives in training to balance the classes.

More on Protein structure prediction