The development of disease and the preservation of health can be analysed in terms of both software and hardware. The software approach, based on information theory, allows biologists to narrow down the range of possibilities and frame hypotheses to direct epidemiological and laboratory based research. These ideas follow from the basic principles which apply to all systems that process information: (i) finite capacity, (ii) finite probability of error, (iii) the error rate rises with age and (iv) complex systems need a high level of redundancy.

The relationship between probability of survival and the number of deleterious mutations in the genome is investigated using three different models of highly redundant systems that interact with a threatening environment. Model one is a system that counters a potentially lethal infection; it has multiple identical components that act in sequence and in parallel. Model two has many different overlapping components that provide three-fold coverage of a large number of vital functions. The third model is based on statistical decision theory: an ideal detector, following an optimum decision strategy, makes crucial decisions in an uncertain world. The probability of a fatal error is reduced by a redundant sampling system, but the chance of error rises as the system is impaired by deleterious mutations. In all three cases the survival profile shows a synergistic pattern in that the probability of survival falls slowly and then more rapidly. This is different than the multiplicative or independent survival profile that is often used in mathematical models. It is suggested that a synergistic profile is a property of redundant systems.

Model one is then used to study the conservation of redundancy during sexual and asexual reproduction. A unicellular haploid organism reproducing asexually retains redundancy when the mutation rate is very low (0001 per cell division), but tends to lose high levels of redundancy if the mutation rate is increased (001 to 01 per cell division). If a similar unicellular haploid organism has a sexual phase then redundancy is retained for mutation rates between 0001 and 01 per cell division. The sexual organism outgrows the asexual organism when the above mutation rates apply. If they compete for finite resources the asexual organism will be extinguished. Variants of the sexual organism with increased redundancy will outgrow those with lower levels of redundancy and the sexual process facilitates the evolution of more complex forms. There is a limit to the extent that complexity can be increased by increasing the size of the genome and in asexual organisms this leads to progressive accumulation of mutations with loss of redundancy and eventual extinction. If complexity is increased by using genes in new combinations, the asexual form can reach a stable equilibrium, although it is associated with some loss of redundancy. The sexual form, by comparison, can survive, with retention of redundancy, even if the mutation rate is above one per generation.

The conservation and evolution of redundancy, which is essential for complexity, depends on the sexual process of reproduction.

A model is described of a highly redundant complex organism that has overlapping banks of genes such that each vital function is specified by several different genetic systems. This generates a synergistic profile linking probability of survival to the number of deleterious mutations in the genome. Computer models show that there is a dynamic interaction between the mean number of new deleterious mutations per generation (X), the mean number of deleterious mutations in the genome of the population (Y) and percentage zygote survival (Zs). IncreasedX leads to increasedY and a fall in Zs but it takes several generations before a new equilibrium is reached. If sexual attraction is influenced by the number of deleterious mutations in the genome of individuals thenY is reduced and Zs increased for any given value ofX. This fall inY and rise in Zs is more marked in polygamous than monogamous mating systems. The model is specified such that deleterious mutations can occur without any observable or measurable effect on function. Thus sexual selection, in this organism, for low levels of deleterious mutations cannot be based on assessment of performance. Instead it is based on a simple symmetrical surface pattern that is flawlessly reproduced by organisms with no deleterious mutations, but is less than perfect, and therefore less attractive, if genetic systems have been deleted. A complex vital task requires a system with a high level of redundancy that acts so that the loss of one component has no observable effect and therefore cannot be used for sexual selection. The reproduction of a beautiful surface pattern also requires a low error, high redundancy genetic system; however, in this case there is advantage if a single deleterious mutation produces a recognisable change. This leads to the conclusion that sexual selection and sexual attraction should be based on beauty rather than utility, and explains the common observation in nature that it is the most beautiful that survive.