Thank you for developing such a valuable tool—our velocity-related analyses heavily depend on it.
We recently started incorporating ambient RNA removal (e.g., using CellBender) into our scRNA-seq analysis pipeline. However, we noticed that when running velocyto run10x, the loom file generated contains counts before ambient RNA removal, which poses a challenge.
To address this, here’s the workaround we’re considering:
Run velocyto on all barcodes using the BAM file generated by Cell Ranger as input. We do this because CellBender requires information on empty droplets to effectively remove ambient RNA.
Extract the spliced, unspliced, and ambiguous counts from the Velocyto results.
Run CellBender to remove ambient RNA.
Convert the CellBender-processed results into a loom file for subsequent velocity analysis.
Could you please provide your thoughts on whether this is a valid approach? I’m concerned that this process might violate some of Velocyto’s prerequisites or assumptions. Any guidance or suggestions you could offer would be greatly appreciated!
Thank you for developing such a valuable tool—our velocity-related analyses heavily depend on it.
We recently started incorporating ambient RNA removal (e.g., using CellBender) into our scRNA-seq analysis pipeline. However, we noticed that when running
velocyto run10x, the loom file generated contains counts before ambient RNA removal, which poses a challenge.To address this, here’s the workaround we’re considering:
Could you please provide your thoughts on whether this is a valid approach? I’m concerned that this process might violate some of Velocyto’s prerequisites or assumptions. Any guidance or suggestions you could offer would be greatly appreciated!
Thank you!