Xdet - MTreedet
Xdet implements two methods for detecting residues responsible for functional especificity in multiple sequence
alignments. These kind of positions, presenting a family-dependent (or function-dependent) conservation pattern, complement the fully-conserved positions as predictors of functionality. They are usually related to functional specificity.
The first method is the "mutational behaviour (MB) method", previously implemented in the MTreedet program.
This method compares the mutational behavior of a position with the mutational behavior of the whole alignment
with the idea that positions showing a family-dependent conservation pattern would have similar mutational behaviors
as the whole family. This method is described in detail in:
The second method is an upgraded version of the MB-method which incorporates the possibility of using an external arbitrary functional classification (in the form of pairwise protein functional similarities) instead of relying on the one implicit in the alignment. Such possibility is intended for cases where a phylogeny/function disagreement is suspected. This method is described in:
A possible measure of protein functional similarity is that representing similarities of "interaction contexts". Fed with that information, Xdet would detect positions responsible for interaction specificity:
If matrices with similarities of physico-chemical properties are used for representing the mutational behaviour of the alignment positions (instead of a substitution matrix) it is possible to detect the physico-chemical properties more related to the changes in specificity, what provides clue on the molecular mechanism behing that specificity:
- Antonio del Sol, Florencio Pazos & Alfonso Valencia.
(2003). Automatic Methods for Predicting Functionally Important
Residues. Journal of Molecular Biology. 326(4):1289-1302.
Xdet is freely available for academic users. Other users, please contact Dr. Florencio Pazos in the address below.
Xdet is now included in the JDet package. In that way, you can run Xdet and inspect its results in an interactive graphical user interface. If, for whatever reason, you want a single command-line version of Xdet not include in JDet's distribution, please contact Florencio pazos in the address below.
The documentation of the program, including information for using it and interpreting the results is available in this
Please, cite the references above when reporting any data obtained using this program.
- Two homology matrices in Maxhom format:
The program can also read any other aminoacid similarity matrix in raw format (see the README file above).
- JDet -
Multiplatform software for the interactive calculation and visualization of function-related conservation patterns in multiple sequence alignments and structures. It includes Xdet and other methods for detecting SDPs and conserved residues in protein alignments.
- MCdet. Another supervised method for detecting specificity-determining positions from multiple sequence alignments.
- Python script for generating the matrix of interaction-based functional similarities for the proteins in a multiple sequence alignment (MSA). More information on the script's code and Pitarch et al. (2020) above.
- Some other servers for predicting functional residues:
Evolutionary Trace @Cambridge. Evolutionary Trace method.
SDPpred @Russia. SDPpred method for predicting functional specificity determining positions (SDP's).
- ConSurf @Israel.
Evolutionary weighted conservation mapped on 3D structures.
- ConSeq @Israel.
Evolutionary weighted conservation combined with PREDICTED solvent accessibility.
Detection of multiple sequence alignment regions that conserve the overall phylogeny of the complete family.
- Hotpatch. Prediction of functional patches in protein 3D structured. Based in the search for surface patches of unusual characteristics.
- Multi-RELIEF. Another method for localting specificity-determining positions in multiple sequence alignments. It can also optionally use 3D information, if available.
- INTREPID. Another server for locating group-dependent conserved positions (as well as fully conserved positions) in multiple sequence alignments.
- FINDSITE - Prediction of ligand binding sites in structures (or threading-based predicted structures). Based on conservation of ligand binding sites in remotely related proteins.
- Specific for protein-protein interaction sites:
- ISPRED @Bologna.
Prediction of protein interaction sites from sequence alignments and 3D structure.
- SHARP2. Protein-Protein interaction sites prediction from 3D structures using Patch Analysis.
- ODA. Protein-Protein interaction sites prediction from 3D structures
based on the search for surface patches with favorable energy change when buried upon protein binding.
- SPPIDER. Protein-Protein interaction sites prediction from 3D structure based on solvent accesibility and other features.
- meta-PPISP. Metaserver for predicting protein-protein interaction sites in 3D structures. It combines the results of 3 methods.
- ISIS. Prediction of protein-protein binding sites for SINGLE SEQUENCES. It can optionally uses 3D info if available.
- Based on low-level structural features predictable from sequence and related to interactions.
- DISOPRED2 server. Prediction of natively un-structured regions.
- COILS server. Prediction of coiled-coil regions.
- ANCHOR - Prediction of Protein Binding Regions in Disordered Proteins.
- Review on co-evolution based methods for extracting information from multiple sequence alignments, including SDP-oriented methods.
Dr. Florencio Pazos Cabaleiro.
Computational Systems Biology Group.
Centro Nacional de Biotecnologia (CNB-CSIC)
Campus U. Autonoma.
Cantoblanco. 28049 Madrid. Spain
Tlf. +34.915854669. Fax. +34.915854506