Previously, log1p normalization was used for the ATAC data in the preprocess tutorial. This raises downstream errors in both the clustering and the visualization pipelines, as those expect a signac_norm layer, which is created when using TFIDF normalization with flavor signac for ATAC normalization.
I decided to adapt the ATAC normalization in the preprocess step instead of in both downstream pipelines, but I'd love to hear opinions on whether this is the best way to resolve this issue.
Previously,
log1pnormalization was used for the ATAC data in the preprocess tutorial. This raises downstream errors in both the clustering and the visualization pipelines, as those expect asignac_normlayer, which is created when usingTFIDFnormalization with flavorsignacfor ATAC normalization.I decided to adapt the ATAC normalization in the preprocess step instead of in both downstream pipelines, but I'd love to hear opinions on whether this is the best way to resolve this issue.
Closes #26.