Title | ||
---|---|---|
Comparative Assessment Of Data Sets Of Protein Interaction Hot Spots Used In The Computational Method |
Abstract | ||
---|---|---|
It seems that every biological process involves multiple protein-protein interactions. Small subsets of residues, which are called "hot spots", contribute to most of the protein-protein binding free energy. Considering its important role in the modulation of protein-protein complexes, a large number of computational methods have been proposed in the prediction of hot spots. In this work, we first collect lots of articles from 2007 to 2014 and select nine typical data sets. Then we compare the nine data sets in different aspects. We find that the maximum number of interface residues used in the previous work is 318, which can be selected as the fittest training data set used in predicting hot spots. At last, we compare and assess the features used in different works. Our result suggests that accessibility and residue conservation are critical in predicting hot spots. |
Year | Venue | Keywords |
---|---|---|
2014 | INTELLIGENT COMPUTING IN BIOINFORMATICS | proteins-protein interaction, hot spots, computational method, training data |
Field | DocType | Volume |
Training set,Data set,Pattern recognition,Computer science,Hotspot (geology),Artificial intelligence | Conference | 8590 |
ISSN | Citations | PageRank |
0302-9743 | 0 | 0.34 |
References | Authors | |
8 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yun-Qiang Di | 1 | 0 | 0.68 |
Changchang Wang | 2 | 0 | 0.34 |
Huan Wu | 3 | 0 | 0.68 |
Xinxin Yu | 4 | 0 | 0.34 |
Junfeng Xia | 5 | 144 | 20.14 |