how to create input?

[zyh4482]

SICILIAN postprocessing all sicilian_called_splice_juncs.tsv to four files namely runname.pq, runname.tsv, runname_GLM_outputs_consolidated.txt and runname_with_postprocessing.txt for a 10x dataset.

May I ask:

1. what do these files mean?
2. I found only runname_with_postprocessing.txt has the called column. The format of runname.pq and runname.tsv looks the same with small test data. Could you please specify which files should be taken as input for SpliZ?
3. According to SICILIAN pipeline, the cell and tissue annotation is not performed during postprocessing. Moreover, the columns in small.tsv is different from the output of SICILIAN postprocessing. It seems that I need to recreate the data by annotating all cells, removing columns that do not exist in small.tsv and sorting the columns. How do you create small.tsv?

Thanks.

[MEFarhadieh]

Dear @zyh4482 as my experience with SpliZ (I'm not in SpliZ team), you should set runname_with_postprocessing.txt without remove any columns as an input file. Also you have to annotate runname_with_postprocessing.txt with condition and cell type as grouping levels or may any extra Information you want. I did this by merging annotation data frame with runname_with_postprocessing according to the barcodes in R.

[zyh4482]

@katze99 Thank you so much! If I have both tissue type and cell type annotations and I used dplyr or other packages to do data cleaning and processing, do you mean runname_with_postprocessing.txt will have two additional columns with tissue type and cell type information in this input file eventually? Then the annotated file can be used as input for SpliZ?

[MEFarhadieh]

@zyh4482 Yes exactly. That file would be fine with SpliZ.