Computer simulation of DNA supercoiling.

We treat supercoiled DNA within a wormlike model with excluded volume. A modified Monte Carlo approach has been used, which allowed computer statistical-mechanical simulations of moderately and highly supercoiled DNA molecules. Even highly supercoiled molecules do not have a regular shape, though with an increase in writhing the chains look more and more like branched interwound helixes.

Neural network models for promoter recognition.

The problem of recognition of promoter sites in the DNA sequence has been treated with models of learning neural networks. The maximum network capacity admissible for this problem has been estimated on the basis of the total of experimental data available on the determined promoter sequences. The model of a block neural network has been constructed to satisfy this estimate and rules have been elaborated for its learning and testing.

[The ability of neuronal networks to generalize using an induction method].

Inductive generalization means the ability of a neural network to learn a given algorithm using incomplete information about it. A consideration based on the information theory leads to a simple equation connecting characteristics of the network with those of the algorithm to be learned. The main conclusion is that the most efficient generalization is achieved on the networks with minimal complexity sufficient for realization of the algorithm under consideration.

Variance of writhe for wormlike DNA rings with excluded volume.

We have calculated the variance of the equilibrium distribution of a circular wormlike polymer chain over the writhing number, less than (Wr)2 greater than, with allowance for the excluded volume effects. Within this model the less than (Wr)2 greater than value is a function of the number of Kuhn statistical segments, n, and the chain diameter, d measured in Kuhn statistical lengths, b. Simulated DNA chains varied from 200 to 10,000 base pairs and the d value varied from 0.

On the ability of neural networks to perform generalization by induction.

The ability of neural networks to perform generalization by induction is the ability to learn an algorithm without the benefit of complete information about it. We consider the properties of networks and algorithms that determine the efficiency of generalization. These properties are described in quantitative terms.