Follow Binomial lab on F6S

A deep learning method for predicting antibody-antigen interactions based on sequence information

Excellent! Now let's get to work!
We are looking forward to receiving your data, if you haven't sent it to us before. Stay in touch so that our specialists can contact you for more information!

Data platform for Antibody-Antigen calculations:
Antigen-Fab-Fab

Point mutations by somatic hypermutation increase variation into the variable region and occur a million-fold more frequently than other genetic mutations.

Somatic Hypermutation

SHM facilitates the progressive increase in antibody affinity against the antigen known as affinity maturation (Figure 4). Some immunoglobulin mutants bind antigens better than others, while some mutants produce non-productive rearrangements. Mutations in the framework regions of the variable domain tend to be selected against as they don’t enhance antigen-binding and alter the basic antibody structure. Mutations that enhance antigen-binding tend to be clustered in the CDR regions.
Strong-affinity binders are selected for and proliferate and mature into antibody-secreting cells, whereas lower affinity clones are eliminated by apoptosis.
Affinity maturation increases antibody activity through multiple rounds of somatic hypermutation and selection in the germinal center.

The structure of a typical antibody molecule

As an antibody, structures are prepared for calculation that contain directly flexible chains that are in direct contact with the Antigen.
Below are the recommended schemes of the biological complex that speed up the operation of the calculation server.

To select flexible antibody chains, it is necessary to modify the contact sites, both on the antibody and antigen sides, as desired.
Preparation of Antibodies for calculations.
Large Antibody structures will be automatically moved to the end of the queue.
Preparation of the calculation complex as a whole:
Obtain a complete set of physical interaction parameters for Antigen-Fab2 that have contact through interacting domains.
  • An example of the PDB structure is given below on the page.
  • An example of the obtained calculated data, correlation with experiment and the use of machine learning methods, see below on the page.
You can introduce mutations into one PDB at once, up to 4 mutations at the same time.
It is desirable that the mutations do not occur one after another, but are spaced and separated by other amino acid residues.
The server does not SEARCH FOR THE CONTACT AREA;
it is assumed that the structure's PDB already contains this spatial information.

  • To obtain more accurate results, it is necessary to separately analyze mutations in the heavy chain from mutations in the light chain, and separately analyze their simultaneous presence in all chains.

Input data

  • Load PDB file N1 for first protein structure,
  • select amount of mutaions,
  • Serial number of this a.a. in PDB file,
  • new indicates of a.a.
  • Load PDB file N2 for second protein structure
  • select amount of mutaions,
  • Serial number of this a.a. in PDB file,
  • new indicates of a.a.
  • Load PDB file N3 for next protein structure
  • select amount of mutaions,
  • Serial number of this a.a. in PDB file,
  • new indicates of a.a.
Job's Name
Enter your e-mail for results
InPut PDB N1 of protein srtucture
How much mutations in peptide N1 do you need?
The serial number of the substitution (mutation) of the amino acid residue according to the PDB file, if there are several substitutions, then write separated by a space
Indicate the new names of amino acid residues after their replacement for the missense mutation
InPut PDB N2 of protein srtucture
How much mutations in peptide do you need?
InPut pdb file N3
 
Example of PDB file structure:
Antibody Fab EGFR
Made on
Tilda