QeneQuant raw reads vs normalized

[davidsanin]

Hi! Thanks for the cool tool!
When I start from estimated read counts from a different pipeline (STAR > featureCounts) should I provide Taiji with raw reads or should I normalise them somehow before running?

My input looked like this and everything went smoothly.

ATAC-seq:
  - id: Naive_ATAC
    group: Naive
    replicates:
      - rep: 1
        files:
          - path: /ATAC/Naive_1.short.cleaned.bam
            tags: ['PairedEnd']

RNA-seq:
  - id: Naive_RNA
    group: Naive
    replicates:
      - rep: 1
        files:
          - path: /RNA_counts/Naive_1.tsv
            tags: ['GeneQuant']

Also my ATAC BAM files were pre-filtered (removed duplicates and longer fragments). Should I add the "Filtered" tag to it?
I just wanted to make sure I was not missing something! Thanks!!

[kaizhang]

Normalized gene expression should be used. Adding the "Filtered" tag won't affect the result but will speed up the processing.

[davidsanin]

Thanks so much for your quick response!
Also, would you mind clarifying some of the meaning of the columns in the network output files?

in edges_binding, what is "affinity"?

# A tibble: 3,478,330 × 8
   `:START_ID`           `:END_ID` chr   `start:int` `end:int` annotation affinity `:TYPE`
   <chr>                 <chr>     <chr>       <dbl>     <dbl> <chr>         <dbl> <chr>  
 1 NFAC4_MOUSE.H11MO.0.C UBL5      chr9     20642715  20642725 promoter      0.543 BIND   
 2 ELK3_MOUSE.H11MO.0.D  UBL5      chr9     20642721  20642733 promoter      0.540 BIND   
 3 RUNX3_MOUSE.H11MO.0.A UBL5      chr9     20642725  20642737 promoter      0.532 BIND 

in edges_combined, how is this weight calculated?

# A tibble: 2,164,825 × 4
   `:START_ID`           `:END_ID` weight `:TYPE`          
   <chr>                 <chr>      <dbl> <chr>            
 1 COT2_MOUSE.H11MO.2.B  UBL5      0.0592 COMBINED_REGULATE
 2 CTCF_MOUSE.H11MO.0.A  UBL5      0.0612 COMBINED_REGULATE
 3 CTCFL_MOUSE.H11MO.0.A UBL5      0.0615 COMBINED_REGULATE

In nodes, the expressionZscore is centered how?

# A tibble: 25,321 × 3
   `geneName:ID` expression expressionZScore
   <chr>              <dbl>            <dbl>
 1 0610005C13RIK      0.898            4.98 
 2 0610006L08RIK      0.333            0.316

Thanks again!!!