This chapter describes the behavior of the immune system from an informationprocessing perspective. It reviews a series of projects conducted at the University of New Mexico and the Santa Fe Institute, which have developed and explored the theme "immunology as information processing." The projects cover the spectrum from serious modeling of real immunological phenomena, such as crossreactive responses in animals and the generation of diversity, to computer science applications, especially the attempt to develop an immune system for computers to protect them against viruses, intrusions, and other malicious activities. In each project, we have used an approach with the following steps: (1) Identify a specific mechanism that appears to be interesting computationally, (2) write a computer program that implements or models the mechanism, (3) study its properties through simulation and mathematical analysis, and (4) demonstrate its capabilities, either by applying the ...

The immune system uses many strategies to generate its enormous repertoire of diverse antibodies, but their relative importance is not understood. Here we address the contribution of antibody gene libraries to the antibody repertoire. We introduce a general framework, in which we can study many antibody-pathogen matching rules, including the widely-used shape-space model (Perelson and Oster, 1979). We use the genetic algorithm as a model of evolution to investigate the type of antibody repertoires that might evolve in relation to a given pathogenic environment. For the antibody/pathogen matching rules that we studied, the scaling relation between fitness and the size of the evolved antibody library is only a shifted variant of the scaling relation that we obtain with random libraries of the same size. We discuss how our results compare to the antibodies that are expressed in newborns, and we discuss the implications of our results for recent experiments with phage a...

The biological immune system exhibits powerful information processing capabil-ities, and therefore is of great interest to the computer scientist. A rapidly expanding research area has attempted to model many of the features inherent in the natural im-mune system in order to solve complex computational problems. This thesis examines the metaphor in detail, in an effort to understand and capitalise on those features of the metaphor which distinguish it from other existing methodologies. Two problem domains are considered — those of scheduling and data-clustering. It is argued that these domains exhibit similar characteristics to the environment in which the biological immune system operates and therefore that they are suitable candidates for application of the metaphor. For each problem domain, two distinct models are developed, incor-porating a variety of immunological principles. The models are tested on a number of artifical benchmark datasets. The success of the models on the problems considered confirms the utility of the metaphor. i

In severe combined immunodeficient (scid) mice, V(D)J recombination is severely impaired due to a recessive mutation ( scid). Thus, we were surprised to find in this study that V�1–J�1 rearrangement is routinely detectable in scid fetal liver, adult bone marrow, and spleen in the apparent absence of completed VH–DJH and V�–J� rearrangements. Particularly surprising, we found the level of V�1–J�1 rearrangement in scid fetal liver to be comparable to that in fetal liver of wild-type mice. The majority of scid V�1–J�1 rearrangements contained abnormal deletions at the VJ junction, consistent with the known effect of scid.

This dissertation is approved, and it is acceptable in quality and form for publication on microfilm: Approved by the Dissertation Committee: Accepted:

Mechanisms of B cell activation and

As predicted by the amino acid sequence, the purified protein coded by Schizosaccharomyces pombe SPAC2F7.06c is a DNA polymerase (SpPol4) whose biochemical properties resemble those of other X family (PolX) members. Thus, this new PolX is template-dependent, polymerizes in a distributive manner, lacks a detectable 30!50 proofreading activity and its preferred substrates are small gaps with a 50-phosphate group. Similarly to Polm, SpPol4 can incorporate a ribonucleotide (rNTP) into a primer DNA. However, it is not responsible for the 1–2 rNTPs proposed to be present at the mating-type locus and those necessary for mating-type switching. Unlike Polm, SpPol4 lacks terminal deoxynucleotidyltransferase activity and realigns the primer terminus to alternative template bases only under certain sequence contexts and, therefore, it is less errorprone than Polm. Nonetheless, the biochemical properties of this gap-filling DNA polymerase are suitable for a possible role of SpPol4 in non-homologous end-joining. Unexpectedly based on sequence analysis, SpPol4 has deoxyribose phosphate lyase activity like Polb and Poll, and unlike Polm, suggesting also a role of this enzyme in base excision repair. Therefore, SpPol4 is a unique enzyme whose enzymatic properties are hybrid of those described for mammalian Polb, Poll and Polm.