Welcome to Qgrid:
Detection of charged and hydrophobic clusters in proteins


This server provides cluster tree diagrams for proteins based on the charges on the atoms or hydrophobicity of a residue. Euclidean distances between charged atoms (or CA atoms of hydrophobic residues) are calculated and cluster tree diagrams are generated using hierarchical clustering method. These diagrams can be used to visually inspect hydrophobic and charged regions in proteins quickly. More help is available Here .

Reference: Shandar Ahmad and Akinori Sarai, Nucl. Acid Res. 32 (2004) W104-107
To get a Qgrid plot of your protein, please upload a co-ordinate file in PDB format (Maximum file size is 1MB).

To get a Cluster tree diagram of a protein in the Protein Data Bank, enter its PDB code here.

Please choose the protein chain, you like to cluster.
(Use the chain name originally used in the PDB file. It is case sensitive).
Make a choice of cluster type:


Which charges to cluster (Not valid for hydrophobic clusters)?

Choose a Charge cutoff:
(for hydrophobic clusters, it means the minimum number of CA atoms of a hydrophobic residue within a grid). Grids having charges with magnitude less than this value will left out of clustering.

Choose a grid size (Angstrom units):
All space will be divided into cubic grids of this dimension. Charges on atoms located in these grids will be added and these grids will be clustered together based on their separation in space. For hydrophobic clusters, all atoms will be assigned zero hydrophobicity except C-alpha atoms of the hydrophobic residues, which will be assigned 1.0 hydrophobicity. This value of 1.0 will in that case be treated equivalent to the the electric charge in charged clusters.

Choose a method of tree joining (This means how the distance between the two clusters of grids should be calculated. If you expect a cluster with speherical, cubic or more symmetric shapes, choose the default Average Linkage. If the clusters are more like an alpha helix, linear or too oblate, a single linkage may be better.
See help file ).