Mutual information without the influence of phylogeny or entropy dramatically improves residue contact prediction

By S. Dunn, Lindi M. Wahl, G. Gloor. 2008

493 citations recorded by the source corpus.

Subfields: Materials Chemistry, Molecular Biology, Computational Biology & Bioinformatics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Materials Chemistry

1. David T. Jones
2. Sergey Ovchinnikov
3. Martin Weigt
4. Arne Elofsson
5. Badri Adhikari
6. Johannes Söding
7. Jianlin Cheng
8. David Baker

In Molecular Biology

1. David T. Jones
2. Martin Weigt
3. Sergey Ovchinnikov
4. Yang Zhang
5. Jinbo Xu
6. Arne Elofsson
7. Debora Marks
8. David Baker

In Computational Biology & Bioinformatics

1. David T. Jones
2. Martin Weigt
3. Sergey Ovchinnikov
4. Jinbo Xu
5. Yang Zhang
6. Debora Marks
7. Arne Elofsson
8. David Baker

Citations in the corpus

• MSA Transformer (2021, methodology): “We also compare against Potts models using the APC-corrected (Dunn et al., 2008) Frobenius norm of the coupling matrix computed on the MSA (Kamisetty et al., 2013).”
• Protein 3D Structure Computed from Evolutionary Sequence Variation (2011, background): “In addition, the principal technical causes of confounding residue correlations are transitivity of correlations, statistical noise due to small numbers and phylogenetic sampling bias in the set of sequences assembled in the protein family [12,13,14,15].”
• Direct-coupling analysis of residue coevolution captures native contacts across many protein families (2011, background): “The problem has been discussed extensively in the literature (10, 23-26).”
• Robust and accurate prediction of residue–residue interactions across protein interfaces using evolutionary information (2014, methodology): “We correct for differences in s ij due to sequence variability at different positions using the row and column averages of these values: where brackets indicate averages taken over the indices outside the brackets in a manner similar to that of Average Product Correction (APC, Dunn et al., 2008 ).”
• MetaPSICOV: combining coevolution methods for accurate prediction of contacts and long range hydrogen bonding in proteins (2014, background): “The next set of input features (six in total) provide coevolution information for each pair of alignment columns, and are as follows: mutual information, normalized mutual information (Dunn et al., 2008), mean contact potential, PSICOV score, mfDCA score and CCMpred score.”
• A new ensemble coevolution system for detecting HIV-1 protein coevolution (2015, methodology): “As an example, CNPR predicted a true positive coupling A45 + E74 (Euclidean distance: 5.69 Å) within the subunit interaction interfaces involving with 0 1% 2% 3% 4% APC ASC BN CMPro 0 1% 2% 3% 4% CTMP CoMap Complementary DCA 0 1% 2% 3% 4% DNcon GREMLIN Interdependency LogR 0 1% 2% 3% 4% MI MIBP Muta”
• Inferring interaction partners from protein sequences using mutual information (2018, methodology): “For structure prediction, these global statistical models outperform 333 the use of MI [4, 6], and have yielded major progress in the field, despite the promising results 334 obtained by MI-based methods implementing corrections for background MI [37].”
• From Principal Component to Direct Coupling Analysis of Coevolution in Proteins: Low-Eigenvalue Modes are Needed for Structure Prediction (2012, unknown): “The direct use of correlations for discovering structural constraints such as residue-residue contacts in a protein fold has, unfortunately, remained of limited accuracy [5,6,9,11,13,16].”