Mouse Cell Atlas reads are aligned to mm9 whereas basenji predicts from mm10

[GrossTor]

The Cross-species regulatory sequence activity prediction project uses bigwig tracks from the mouse cell atlas. Those were generated from ATACseq reads that were aligned to the mm9 genome assembly. However, basenji tries to predict those tracks from the sequence of the mm10 genome assembly (other mouse tracks e.g. from Fantom5 are aligned to mm10 as well).

Could you explain how this discrepancy was resolved? I see multiple solutions:

• use Crossmap to map the bigwig files to mm10
• retrieve the fastq files from the Mouse Cell Atlas and generate mm10 aligned bigwigs with the basenji data pipeline
• align all other mouse tracks to mm9 and only work on mm9

What would you recommend?

[davek44]

That's correct. Crossmap seems good. I converted to bedgraph, then used UCSC liftover.

[GrossTor]

Makes sense. Thank you!