Analysis of an HIV-infected cohort followed for as long as 15 years after seroconversion.

Data from a cohort of 62 HIV-positive individuals with hemophilia or von Willebrands disease infected for a maximum period of 15 years were analyzed. The relation between CD4+ and total lymphocyte counts and their rate of decline was analyzed with respect to age at seroconversion, time of seroconversion, and development of disease and subsequent death. As expected, the CD4+ and total lymphocyte population decline correlated with increased probability of disease and death. The patients fell into two distinct categories with respect to this decline: those whose cell count declined steadily (single slope) and those whose cell count remained steady or decreased very slowly for a variable period and then declined sharply (double slope). Within this cohort, the presence of a double slope appears to indicate a poorer prognosis, as 9 of 18 of the patients who have died showed this pattern, whereas only 6 of 42 of the remaining patients have this pattern even though more than half of this group have CD4+ lymphocyte counts < 0.2 x 10(9)/L. In addition, the ratio of CD4+ lymphocyte count to total lymphocyte count decreased with increasing cumulative frequency of the cumulative incidence of disease and death and the overall probability of death in this cohort was lower than expected, being 30% 12 years after seroconversion.

Correlations between the coding and non-coding regions in DNA.

In this paper various aspects of codon usage and k-tuple correlations in the DNA are compared. It is shown that the correlation structures of the coding and the non-coding regions are very similar and that codon usage is reasonably specific for large groups of organisms. These results suggest that the origin of codon usage is related to the origin and structure of the DNA.

Long-range doublet correlations in DNA and the coding regions.

Long-range two-body correlations in a DNA sequence should in theory approach a constant value very rapidly with increasing value of the correlation length. It is shown that for most DNA sequences, the long-range correlations exhibit oscillations superimposed on the constant background. These oscillations persist for very large correlation lengths. The oscillations are shown to be three-point cycles and are related to the coding regions in the DNA. A method for discovering the coding regions in DNA sequences is presented. The limitations of the method are discussed.

A model of quantitative traits under frequency-dependent balancing selection.

We describe a computer model that stimulates a combination of stabilizing and frequency-dependent selection acting on a quantitative character determined by several loci. The results correspond to many features of natural variations at both the phenotypic and genotypic levels. The model is robust, and its results are not strongly dependent either on the nature and shape of the function describing the stabilizing selection, or on the precise form of frequency dependence, except near the extrema. It suggests a mechanism for the maintenance of large amounts of variability, and shows a relation between population size and heterozygosity roughly corresponding to that found in nature. In this respect it is unlike the purely neutral model.

Mutational order: a major stochastic process in evolution.

Computer simulations in which selection acts on a quantitative character show that the randomness of mutations can contribute significantly to evolutionary divergence between populations. In different populations, different advantageous mutations occur, and are selected to fixation, so that the populations diverge even when they are initially identical, and are subject to identical selection. This stochastic process is distinct from random genetic drift. In some circumstances (large populations or strong selection, or both) mutational order can be greatly more important than random drift in bringing about divergence. It can generate a 'disconnection' between evolution at the phenotypic and genotypic levels, and can give rise to a rough 'molecular clock', albeit episodic, that is driven by selection. In the absence of selection, mutational order has little or no effect.

Evolution of resistance with sequential application of insecticides in time and space.

The effect, on the evolution of resistance, of alternating two unrelated insecticides in space or in time (or both) is studied. Transient polymorphism is shown to occur under certain conditions of mating, selection and migration. In some situations, the transient polymorphism can show a sharp decline before the alleles recover to fixation. Alternating a single insecticide in space, and in space and time, is also considered. Neither alternation in space nor in time shows any advantage with regard to delaying the onset of resistance. The most promising mode is to alternate the presence and absence of a single insecticide in both space and time, especially if it is applied at the larval stage and if some form of biological control is used in the regions where no insecticide is applied.

Frequency-dependent selection, metrical characters and molecular evolution.

Computer models of selection acting on a quantitative character show that a combination of frequency-dependent and stabilizing selection can maintain many polymorphisms among the genes that determine the character. The models also show that the random order of mutations can give rise to selectively driven stochastic effects that are sometimes more important than random genetic drift. They suggest simple explanations for patterns of divergence between populations and species, and for apparent discrepancies between the rates of morphological and molecular evolution. They point towards a selective theory of 'molecular clocks'.

Can phenotypic variability of polygenic traits yield information regarding selection?

The relation between phenotypic variability and heterozygosity for a polyallelic polygenic trait is obtained using computer simulation. It is shown that unlike the di-allelic case, the variance of phenotypic score of the trait can increase or decrease depending on the allele frequency distribution. The relevance of this to selection is discussed.

What can be learnt about selection from gene frequency distribution?

Polymorphism has been studied at the Esterase 6 locus in the Yellow Fever mosquito Aedes aegypti (L.) in laboratory stocks. At least 12 alleles are present, with up to four coexisting in a stock. The allele frequency distribution is quite sharply peaked at a mode of about 0.25. The experimental data are compared with the results of simulation based on two models, one in which the initial global distribution is taken to be the stationary distribution obtained from the neutral model assuming M = 4 mu Ne = 1 and the other in which the initial global distribution is generated from the experimental populations studied. The results suggest that the patterns observed are not likely to arise through random fluctuation of frequencies in neutral alleles, but that some kind of selection maintains polymorphism, either in the wild or in the laboratory, or both.

Postindustrial melanism in the peppered moth.

New data show the geographical pattern of frequency of the melanic morph carbonaria of the peppered moth, Biston betularia, in 1983-84. These frequencies are compared with data from 1952 to 1970. After 20 years of smoke control, the area of high melanic frequency has contracted to the northeast. The change indicates a disadvantage to carbonaria of about 12 percent compared with 20 years ago. Computer simulations, which do not include the assumption of heterozygote advantage, provide a good match to the surface for the period 1952 to 1970, and also the 1983-84 surface. Experiments on visual predation have been criticized as giving unrepresentative estimates of selection but they permit satisfactory simulations to be made.

Evolution of resistance in the presence of two insecticides.

A two-locus model is used to analyze the effectiveness of a mixture of insecticides in delaying resistance, compared to the use of the insecticides singly. The effects of factors such as recombination, effective dominance, initial value of allele frequencies and initial value of linkage disequilibrium are considered. It is shown that the use of mixtures is always more effective in delaying the onset of resistance, often by many orders of magnitude. It is shown that there exists a threshold value of recombination fraction, above which the evolution of resistance is extremely slow. Resistance evolves very rapidly for values of recombination fraction below the threshold. Finally, the relevance of these results on resistance management is discussed.