Percentiles in variance decomposition

Dear all,
unfortunately I didn’t find any information on this in the manual of Dynare.

How can one obtain the variance decomposition for different percentiles regarding bayesian estimation? I think I understand the options of using moments_varendo in the estimation command and stoch_simul after the estimation.

For example did Khan and Tsoukalas (2012) indicate the 10-90th Percentile (Table 2) and Schmitt-Grohe and Uribe (2012) indicate the 5th and 95th Percentile (Table V).

I would be thankful for any advice on that issue.

//Edit:
I found the deciles of moments_varendo in oo_.PosteriorTheoreticalMoments.dsge.VarianceDecomposition.deciles.
And for stoch_simul I could set the parameters to the decile i want to.

After estimation,
oo_.PosteriorTheoreticalMoments.dsge.ConditionalVarianceDecomposition.HPDinf for example contains the lower percentile, while HPDsup contains the upper one. The option mh_conf_sig governs the size of the percentile. By default, it is a 90% one, i.e. 5th and 95th Percentile

Thank you very much.
I think I will use the HPD intervals as you suggest and follow the example of Schmitt-Grohe/Uribe.