Input epitope region in Absolut

[Rajitha-ACOG]

Hi, Nice Work !! Using Absolut I am able to discretize a new antigen. For next step of generating the structures for it, Is it possible that Absolut can consider only the region specified by the user as epitope and not generating all possible structures for all possible epitopes in antigen. I am avoiding to consider only the structure of epitope as input as it is observed that the other regions of the antigen impose constraints in the conformation of the epitope. As my understanding goes, considering the entire antigen can impose constraints on the moves the epitope can have in Absolut. Please correct me if I am wrong.

[petouzeman]

Dear Rajitha, Sorry for delay, I don't get notifications, feel free to mail me to get a faster reply, So you are right, and this is possible! If your antigen is relatively small, you can manually block one by one the positions that are around the other epitopes, such that antibodies can only bind to the epitope of interest. I would do that by hotspot rather than epitope, to let a bit of flanking area free for the antibodies to bind. For doing this, you need the version of Absolut with libraries, and when you see the 3D antigen, you can use the following keyboard commands: x,y,z or X, Y, Z and a green point will move in space. You can move this point iteratively to each surface position you would like to block, and each time type 'O', this will output the position you are at. Then in antigenlib.cpp, when you add your antigen, you add those blocked positions manually. A bit tedious but would work! Tell me if it works, or I can help to find another solution! Best, Philippe

[Rajitha-ACOG]

Thank you so much Philippe. I will try that out.

Thanks, Rajitha

[petouzeman]

Dear Rajirha! Yes, this is possible :) this is the option Absolut getFeatures you give the result of bindings as a text files and he creates plenty of columns with these infos, Hope it helps! Best, Philippe

Le lun. 26 déc. 2022 à 13:22, Rajitha-ACOG @.***> a écrit :

Thank you so much Philippe. I will try that out. I am curious to know if the entire antigen is used, is there a way that I can output the antigen residues interacting with CDRH3 in addition to the energy value (In short, can we have a additional column of antigen residues sequence here #ID_slide_Variant CDR3 Best Slide Energy Structure)

Thanks, Rajitha

— Reply to this email directly, view it on GitHub https://github.com/csi-greifflab/Absolut/issues/11#issuecomment-1365135093, or unsubscribe https://github.com/notifications/unsubscribe-auth/ANHKCUFZRQKJXFSKSL5BMQDWPGEXJANCNFSM6AAAAAASVA2TAA . You are receiving this because you commented.Message ID: @.***>

[Rajitha-ACOG]

Thanks, Philippe. I figured it out later that getFeatures option gives the antigen sequence too.

Of curiousity, I am wondering if Absolut can be used for given TCR CDRs and a peptide:MHC complex.

Thanks, Rajitha

[petouzeman]

Dear Rajitha,

Sorry for delay! So, this is a very good question about the peptide MHC, what we do is only to fold a 11-mer onto a 3D structure for a PDB, so you can decide the meaning of the 11-mer and the 3D antigen. In theory, if you would use the full peptide + MHC structure as antigen, then you could fold CDRH3 from TCRs onto it. The only problem would be, TCR are evolved to bind in a certain region of the pMHC (with different angles depending on the MHC allele/locus I believe), while in Absolut, the 11-mer would bind everywhere, which would be maybe incorrect. What you could do is to restrict possible binding areas to regions where a CDRH3 usually binds on the pMHC surface, and block other positions on the surface of the pMHC, but this could be tedious. So I would say, better to keep for antibody-antigen but there might be some ways around. Hope it helps!

Best,

Philippe

On 30/12/2022 05:17, Rajitha-ACOG wrote:

Thanks, Philippe. I figured it out later that getFeatures option gives the antigen sequence too.

Of curiousity, I am wondering if Absolut can be used for given TCR CDRs and a peptide:MHC complex.

Thanks, Rajitha

[rajkumar1501]

Hello Philippe,

Is there an alternative method to determine the integers corresponding to the respective residues, aside from utilizing the visualization technique? For example, can we use the starting position integer and the observer string to derive the corresponding integer positions of the residues?

Best, Rajkumar

[petouzeman]

Dear Rajkumar, Yes, this is possible, these are vector operations (SULDR is a rotation matrix multiplication to the previous direction). This is not given as command line, but if you are willing to use C++ there are functions available. They are inside Ymir subfolder/package, in the compact.h/cpp class. I suggest you to look at Ymir/dox and to unzip the html doc in zip, then you have more explanations on the functions available Hope it helps!