Feature Request/Support: Support for mixed fragment_single and fragment_paired keys in AnnDataSet

I am attempting to combine Single-end (SE) and Paired-end (PE) ATAC-seq samples into a single AnnDataSet for downstream analysis using snapatac2. The Problem is SE samples store fragment information in adata.obsm['fragment_single'] while PE samples store fragment information in adata.obsm['fragment_paired']. When these are combined into an AnnDataSet, the disparate keys in .obsm lead to issues where downstream functions cannot locate the fragment data because they expect a consistent key across all underlying AnnData objects.

What is the best practice for joint analysis of mixed-end libraries in SnapATAC2? Is there a recommended way to "standardize" these keys before merging (e.g., treating PE as SE) without losing data integrity? Does AnnDataSet support mapping these two different keys to a universal fragment interface?

Thank you for your help!

No files or tests are named. Start by tracing how AnnDataSet combines .obsm entries and how downstream functions locate fragment_single or fragment_paired. Done would require a documented, data-preserving approach for mixed single-end and paired-end libraries, with coverage for merging and downstream analysis.