According to the polaron hybrid DNA model, the initial nucleotide-chain breakage leading to genetic recombination takes place only at the linkage points which define the ends of each polaron. The term dissociation cycle is proposed for the postulated series of events from primary breakage at a linkage point to the final breakdown of unpaired chains. At mutant sites, the mispairing in hybrid DNA may persist and give rise to post-meiotic segregation, or a correction process may operate by which molecular homozygosity is restored. This causes conversion, which may be evident as reciprocal or as non-reciprocal recombination.
The implications of this model are that a crossover occupies a segment of the chromosome, and that conversion is a process which takes place when a mutant site happens to lie within such a segment. Although crossovers appear to be initiated at fixed points outside or at the ends of the genes, they extend into the gene on one side or the other. The negative interference between recombination events over short intervals of the linkage map is attributed to the association between crossing-over and the conversion which is likely to occur at any mutant sites which happen to lie within the crossover. In the same way, non-crossover hybrid DNA can also lead to conversion, and hence to negative interference.
The relevant data on genetic recombination have been found to fit this model, and have led to the following main conclusions:
(1) The polaron may coincide with the cistron, or in some instances may possibly include more than one cistron. As a corollary to this, there appear to be two kinds of cistrons: unipolar, where all the recombination is initiated from the same end, and bipolar, where it is sometimes initiated from one end and sometimes from the other.
(2) In bipolar cistrons there is usually a preponderance of recombination initiated from one end over that from the other. In five genes where the orientation with respect to the centromere is known, two show a preponderance in favour of the proximal end and the other three in favour of the distal end. It seems possible that this asymmetry within the gene may reflect intrinsic differences in the frequencies with which the dissociation of the DNA molecules is initiated at different linkage points.
(3) The hypothesis of a fairly constant amount of newly-synthesized DNA per dissociation cycle, irrespective of how it is distributed along the four templates from a linkage point, leads to a number of predictions for which there is evidence in support. These concern the detailed pattern of crossover and non-crossover hybrid DNA within the gene.
(4) Specific differences in intragenic recombination are attributed to differences in the pattern of DNA synthesis. The predominantly non-reciprocal recombination found within cistrons of Neurospora crassa would be explained if synthesis extends unequally from the linkage point in the two chromatids. The higher frequency of reciprocal recombination found in Aspergillus nidulans is attributed to more equal extension.
(5) Differences between meiotic and mitotic intragenic recombination both in A. nidulans and in Saccharomyces cerevisiae are explained on the supposition that more DNA synthesis per dissociation cycle occurs at mitosis than at meiosis.
(6) Clustering of sites in maps of alleles based on recombination frequencies is attributed to a rather limited range of variation in the lengths of the newly synthesized nucleotide chains.