Molecular markers linked to QTL contributing to agronomic and fibre quality traits would be useful for cotton improvement. We have attempted to tag yield and fibre quality traits with AFLP and SSR markers using F₂ and F₃ populations of a cross between two Gossypium hirsutum varieties, PS56-4 and RS2013. Out of 50 AFLP primer combinations and 177 SSR primer pairs tested, 32 AFLP and four SSR primers were chosen for genotyping F₂ individuals.Marker-trait associations were studied for eight agronomic and five fibre quality traits through simple and multiple regression analysis (MRA) using a set of 92 AFLP polymorphic loci and four SSR markers. Simple linear regression analysis (SLRA) identified 23 markers for eight different traits whereas multiple regression analysis identified 30 markers for at least one of the 13 traits. SSR marker BNL 3502 was consistently identified to be associated with fibre strength. While all the markers identified in SLRA were also detected in MRA, as many as 16 of the 30 markers were identified to be associated with respective traits in both F₂ and F₃ generations. The markers explained up to 41 per cent of phenotypic variation for individual traits. A number of markers were found to be associated with multiple traits suggesting clustering of QTLs for fibre quality traits in cotton.

Most traits of interest to medical, agricultural and animal scientists show continuous variation and complex mode of inheritance. DNA-based markers are being deployed to analyse such complex traits, that are known as quantitative trait loci (QTL). In conventional QTL analysis, F2, backcross populations, recombinant inbred lines, backcross inbred lines and double haploids from biparental crosses are commonly used. Introgression lines and near isogenic lines are also being used for QTL analysis. However, such populations have major limitations like predominantly relying on the recombination events taking place in the F₁ generation and mapping of only the allelic pairs present in the two parents. The second generation mapping resources like association mapping, nested association mapping and multiparent intercross populations potentially address the major limitations of available mapping resources. The potential of multiparent intercross populations in gene mapping has been discussed here. In such populations both linkage and association analysis can be conductted without encountering the limitations of structured populations. In such populations, larger genetic variation in the germplasm is accessed and various allelic and cytoplasmic interactions are assessed. For all practical purposes, across crop species, use of eight founders and a fixed population of 1000 individuals are most appropriate. Limitations with multiparent intercross populations are that they require longer time and more resource to be generated and they are likely to show extensive segregation for developmental traits, limiting their use in the analysis of complex traits. However, multiparent intercross population resources are likely to bring a paradigm shift towards QTL analysis in plant species.