Macdonald S. Morris

Models of intergenerational kin altruism

A model for kin altruism involving multiple generations, in which altruism is directed from daughters toward their mothers, is investigated and compared with models for altruism involving a single generation. In the linear case of constant cost and benefit, the conditions for the evolution of intergenerational altruism are shown to depend on the form of expression of this gene, the criteria by which altruistic evolution is assessed, and the level of penetrance of the gene for altruism; these conditions are usually not found in intragenerational models. Central to the interpretation of these results are demographic factors and aspects of life history strategy that play an important role in intergenerational altruism. The case in which the fitnesses of the altruistic and selfish phenotypes are general functions of the fraction of altruists in the brood is also investigated. We find that the shape of these functions has important effects on the conditions for the evolution of altruism. To these effects we give a natural interpretation in terms of the properties of altruistic behavior.

Algorithms for identifying local molecular sequence features.

Efficient algorithms are described for identifying local molecular sequence features including repeats, dyad symmetry pairings and aligned matches between sequences, while allowing for errors. Specific applications are given to the genomic sequences of the Epstein-Barr virus, Varicella-Zoster virus and the bacteriophages lambda and T7.

Exact formulas for multitype run statistics in a random ordering

Exact formulas are developed for the probability that a random ordering of a fixed collection of letters contains any specified collection of runs. These results are particularly useful for short sequences and other cases where asymptotic formulas cannot be trusted. Computer programs to evaluate these formulas are available.