Reader Comments

Post a new comment on this article

Locus specific and Antigen Presentation Bias

Posted by cmcmurtr on 31 Jan 2013 at 17:36 GMT

This is a very interesting analysis of known epitopes. We have a couple of questions:
1. In figure 2 you looked for a positional bias in predicted ligands across multiple HLA; both A and B. Have you looked for a locus-specific bias (HLA-A, HLA-B, or HLA-C) in known epitope position?
2. The two biases for immune recognition (central position and high conservation) are clearly presented. It seems as if DRiPs can be centrally biased, and at the same time immune recognition is centrally biased. It is then implied that antigen presentation of ligands is centrally biased. If you were to examine naturally processed and loaded host ligands, would you expect to see the same bias in class I MHC antigen presentation?

- William Hildebrand and Curtis McMurtrey, OUHSC

No competing interests declared.

RE: Locus specific and Antigen Presentation Bias

ykimbiology replied to cmcmurtr on 11 Feb 2013 at 19:00 GMT

Thanks for your questions!

Regarding 1: We have not looked at locus-specific bias. In Figure 2, we can see that the individual MHC-specific effects are small to begin with, so we do not expect to see locus-specific biases for the predicted binding affinities. In terms of locus specific bias for known epitopes, we also did not look at that, and in general this is more difficult as the known epitope dataset is already pretty small, so subdividing it further is not optimal.


Regarding 2. We believe that the central bias of epitope position is explainable by effects outside of antigen-processing, namely that the bias is due to repeated exposures to conserved epitopes of homologous antigens and their triggering of memory T-cells generated from earlier exposure. However, you are right, we did not look at eluted ligands, and that would be a good approach to test if antigen processing itself has positional bias. The IEDB is not a great place to look for eluted ligands right now, as many elution experiments characterize cancer antigens or self antigens that are not generally in IEDB scope. But given your groups expertise in eluted ligands, it would be great if you can check for positional bias in your dataset!

- Yohan and Bjoern

No competing interests declared.