QC RNA-seq expression outliers

[christie-ga]

Hi, I have a question regarding the QC of the Gtex eQTL pipeline.

There is one step mentioned on the GtexPortal (https://gtexportal.org/home/documentationPage#staticTextAnalysisMethods) that I would like to use on my data:

"1. RNA-seq expression outliers were identified and excluded using a multidimensional extension of the statistic described in (Wright et al., Nat. Genet. 2014 ). Briefly, for each tissue, read counts from each sample were normalized using size factors calculated with DESeq2 and log-transformed with an offset of 1; genes with a log-transformed value >1 in >10% of samples were selected, and the resulting read counts were centered and unit-normalized. The resulting matrix was then hierarchically clustered (based on average and cosine distance), and a chi2 p-value was calculated based on Mahalanobis distance. Clusters with ≥60% samples with Bonferroni-corrected p-values <0.05 were marked as outliers, and their samples were excluded."

Is there a script or some code available that I could use to perform these steps?

Thank you for your help!