Using Portcullis on large Datasets in connection with Mikado

[cc-prolix]

Hello,

I am using Portcullis to analyze and quantify splice junctions of data generated with HISAT2. I have around 3000 BAM Files and am wondering what the best practice would be if I were using the generated splice junction information as input for Mikado (https://mikado.readthedocs.io/en/stable/). As Portcullis merges the provided BAM files if more than one was provided, I wanted to ask if I should analyze each file separately or all at once (and merging them) if I want to use them for your Mikado Pipeline?

Thank you very much for your help!