There is technology available to identify human helper T cell epitopes. Its a computer-based method for predicting the binding of peptides to human MHC class II molecules. Computer models of MHC class II are generated with reference to known structures solved by X-ray crystallography. Individual peptides are then analysed for a preferred conformation for binding within the peptide binding groove of each MHC molecule. My question is : how these peptides are generated? based on a known antinody sequence or randomly chosen?

I am thinking that peptides are randomly chosen from a library. This method is called: peptide threading and the binding could be comparable to in-vivo peptide-MHC binding.

DD said:

There is technology available to identify human helper T cell epitopes. Its a computer-based method for predicting the binding of peptides to human MHC class II molecules. Computer models of MHC class II are generated with reference to known structures solved by X-ray crystallography. Individual peptides are then analysed for a preferred conformation for binding within the peptide binding groove of each MHC molecule. My question is : how these peptides are generated? based on a known antinody sequence or randomly chosen?

Peptides can be eluted from MHC class I or class II molecules and identified by LC/MS/MS or ESI/MS/MS techniques that can be confirmed by reactive CD8+ T cells or CD4+ T cells. A significant proportion of this information is stored in databases that aid in the design of computer algorithms which have embedded matrices containing coefficients of amino acid frequencies occurring at specific positions. Algorithms such as SYFPEITHI and TEPITOPE are aiding the development of synthetics that would mimic such epitopes. The alternative would be to generate synthetics of overlapping peptides, thus covering the entire length of the protein sequence and confirming biological activity using CD4+ Tcells with a readout such as IFNg ELISPOT .