Hello friends of OPIG,

From my last blopig blog post [link: https://www.blopig.com/blog/2019/10/comparative-analysis-of-the-cdr-loops-of-antigen-receptors/], I summarised our findings that TCR CDRs are more flexible than their antibody counterparts. Because of this observation, we believe that it is more appropriate to represent TCR binding sites using an ensemble of conformations.

In our latest tool, TCRBuilder, we adapted our antibody modelling pipeline ABodyBuilder for TCR structural modelling (see Figure). Akin to antibodies, TCR frameworks are rather conserved. Homology models are found using sequence similarity over the framework region. The CDR templates are, similar to ABodyBuilder, searched by FREAD. However, the novelty here is that if we have observed multiple conformations of the CDR sequence of interest in the PDB, we will return an ensemble of models.

Embedded in the pipeline are the ab initio modelling tool Sphinx, and side-chain modelling programme PEARS, supplied with an updated TCR-specific database.

The ensemble output could potentially be used as multiple starting points for further macromolecular dynamics simulations should the user desires.

Try it out yourself! http://opig.stats.ox.ac.uk/webapps/stcrpred/TCRBuilder/

If you like it, cite us here: Wing Ki Wong, Claire Marks, Jinwoo Leem, Alan P Lewis, Jiye Shi, Charlotte M Deane, TCRBuilder: Multi-state T-cell receptor structure prediction, Bioinformatics, , btaa194, https://doi.org/10.1093/bioinformatics/btaa194