Germline diversity does not contribute to the direct recognition of pathogens

In chapter I used a simple model of gene library evolution to investigate the scaling of the survival probability of an organism with the number of antibodies in its repertoire. I showed that for distributions of the antibody-antigen bond strength that I consider of biological relevance, the survival probability of the organism increases logarithmically, or sublogarithmically, with the size of its germline-encoded antibody repertoire. This suggests that the role of germline-encoded immune receptor genes is not specific recognition of individual pathogens, but rather a coarse-grain encoding of the region of the pathogen universe that the species has encountered in evolution. I showed that such an encoding can be achieved if an individual is confronted with a large fraction of the possible pathogens in its life time. High-affinity recognition of a pathogen would then have to be achieved through fine-tuning of the antigen receptor during an immune response. If the number of pathogens that the organism encounters is comparable with the number of receptors that it can encode in germline, then a static pathogenic environment would result in better learning of the training set. However, if the pathogenic environment changes, the performance of the immune system would be lower on the pathogens that it encountered, but it would be higher on random pathogens. The reason for this behavior, which one might view as robustness, is that, although static pathogenic environments result in specialized libraries, dynamic pathogenic environments induce essentially random antibody libraries.