Comparison of algorithms rapid chain delineation and seriation, for the construction of genetic linkage maps

Abstract

The objective of this work was to evaluate the efficiency for the construction of genetic linkage maps of the algorithms seriation and rapid chain delineation, as well as the criteria: product of adjacent recombination fractions, sum of adjacent recombination fractions, and sum of adjacent LOD Scores, used with the ripple algorithm. A genetic linkage map was simulated containing 24 markers with random distances between them, with an average of 10 cM. Using the Monte Carlo method, 1,000 backcross populations and 1,000 F₂ populations were simulated. The populations comprised 200 individuals each, as well as different combinations of dominant and codominant markers (100% codominant, 100% dominant and mixture containing 50% codominant and 50% dominant). It were also simulated 25, 50 e 75% of missing data. It was observed that both algorithms presented similar performance, and were sensitive to the presence of dominant markers, which makes it difficult to get estimates with good accuracy for both order and distance. Moreover, the algorithm ripple, when applied with the criteria sum of adjacent recombination fractions and product of adjacent recombination fractions, increased the number of correct orders.