count function gives different results for the same total count

[tulintatar]

Hi,

I am trying to use mageck countand testto be able to assess the treatment effect.

I have 6 treatments (T) and 3 controls (C).

The problem is : If I choose all 6 Ts and 3 Cs, and choose 'total' for the normalization when I use mageck count and when I use only 1T and 1 C, I have very different results.

My command when I am using 6 T and 3 C is as follows:

mageck count -l Genome/modified_sgRNA.txt -n totalcount --norm-method total --trim-5 26 --sample-label "day0,day01,day02,day1,day2,day3,day4,day5,day6" --fastq Control1.fastq,Control2.fastq,Control3.fastq Treatment1.fastq,Treatment2.fastq,Treatment3.fastq,Treatment4.fastq,Treatment5.fastq,Treatment6.fastq

My results for GLI4:

GLI4_1  GLI4    3700.2680275656676  57.097543911735315
GLI4_2  GLI4    2848.654291843986   11628.86644335676

When I use 1 T and 1 C:

mageck count -l Genome/modified_sgRNA.txt -n totalcount --norm-method total --trim-5 26 --sample-label "day0,day6" --fastq Control1.fastq Treatment6.fastq

Results for same gene:

GLI4_2  GLI4    0.6597633136094675  0.0
GLI4_1  GLI4    1.9792899408284024  0.0

Why am I seeing different results for the same gene, if I am using the total counts in both of the cases? The total count for day0 and day 6 are the same..

I could not see an explanation of this, and I would appreciate to know what this parameter does.

Thank you very much.