MSIsensor parameters

[Jane M]

Dear all,

I am using MSIsensor for the first time.
As a first test, I launched it on my paired (realigned, recalibrated) bam files using default parameters and got the output files without problem.
I was a bit surprised because I expected a high score for MSI for some samples, while it was ~3.7 only for 2 samples.
So now, I would like to clearly understand and refine, if needed, the parameters, knowing that I am working with human samples.

msisensor scan

• context length -c 5
  What does it represent?
• minimal/maximal homopolymer size -l 5 / -m 50
  Nb of bases forming a repeat unit, between 5-50
• maximal length of microsatellite -s 5
  I am not sure: is it the maximal number of bases forming a repeat unit in a microsatellite. Isn't it small?
• minimal repeat times of microsatellite -r 3
  Minimal number of repetitions to form a microsatellite

msisensor msi

• minimal homopolymer size -l 5 / -p 10
  Why changing the threshold for distribution analysis?
• minimal microsatellites size -q 3
  Is it equivalent to -r in scan mode?
  Again, why changing the threshold for distribution analysis with -s 5?
• maximal microsatellites size -w 40
  There was no threshold for that in scan mode. Why this difference?

Does what I understood seem correct and could you give me some clues about the parameters I did not catch?
Thank you in advance for your help