JoinMap is one of the most widely used software for the construction of genetic maps. It is commercially available and benefits from a highly advanced MS-Windows user interface for data management and analysis, professional support and continued development. JoinMap has two distinct marker order search strategies: the regression mapping and maximum likelihood mapping algorithms. The original regression mapping approach uses a goodness-of-fit statistic strategy to judge, at each step, whether the addition of a marker should be accepted. The maximum likelihood approach was developed to deal with larger datasets and is much faster than the regression mapping approach (Jansen et al. 2001). It maximizes a multipoint likelihood function using Gibbs sampling to estimate multipoint recombination frequencies, and simulated annealing to search marker orders. JoinMap also has a range of other functionality such as testing segregation distortion, analyzing similarity of loci and individuals, comparing two maps with homologous markers, testing heterogeneity, etc. (Stam 1993). Among those a powerful capability of JoinMap is able to integrate data from multiple populations to construct consensus maps. The current software version (V4.1) was released in July 2011.

In switchgrass, JoinMap has been used to construct linkage maps in two studies (Okada et al. 2010; Liu et al. 2012). In both examples, linkage maps were constructed in two steps. Initially, using strict LOD thresholds and a maximum-likelihood module, the markers with high fidelity were grouped into linkage groups to construct a framework map. Secondly, those markers on framework map were fixed to allow more markers to be added on the linkage map by reducing the LOD thresholds. This strategy guaranteed both the accuracy and high coverage of the final full maps (Okada et al. 2010; Liu et al. 2012, 2013b).

Recently, Yang et al. (2013) used the same data as Okada et al. (2010), and reconstructed linkage maps of female and male parents, respectively. The maps included 24 linkage groups for female parent and 21 linkage groups for male parent. They claimed the quality of map construction could be improved by using a new algorithm, which allowed simultaneous estimation of the linkage, genetic interference and preferential pairing factors (Yang et al. 2013).