Best transition model for microsatellites?

[Adam S]

I would like to model microsatellites. These are genomic elements with 2, 3, or 4 bp repeating motifs (e.g., CAG,CAG,CAG). I model them as having a length that can increase or decrease by an integer change in repeat units. Changes of +/- 1 unit are the most common with changes of more than 1 unit occurring at an exponentially lower rate. 

I initially tried the approach in this post, but found it to be too slow to use with large datasets. 

Instead I would like to simply model microsatellites as having ordered transitions unsing the fnOrderedRateMatrix function as outlined here under heading "Asymmetric ordered". Is this model appropriate for this use case? If not, is there any other that you recommend? 

Finally, what do the rescaled and matrixExponentialMethod arguments mean? And do you have a suggestion for how these should be set? 

Thank you very much! 

Adam