EdgeR in CIRI_DE_replicate issue

[andre-gabriel-42]

I am performing differential gene expression analysis using the CIRIquant pipeline. Referring to the CIRIquant cookbook, I successfully arrived into this step:

"Usage 3: Differential expression analysis - Study with biological replicates - Step3: Differential expression analysis"

However, there is an EdgeR error occurring (see below). Can someone explain to me what I may have done wrong?

Command used:

CIRI_DE_replicate --lib library_info.csv --bsj circRNA_bsj.csv --gene gene_count_matrix.csv --out circRNA_de.tsv Log:

[Wed 2021-03-17 22:11:31] [INFO ] Library information: /home/andregabriel/Desktop/CIRI_test/library_info.csv
[Wed 2021-03-17 22:11:31] [INFO ] circRNA expression matrix: /home/andregabriel/Desktop/CIRI_test/circRNA_bsj.csv
[Wed 2021-03-17 22:11:31] [INFO ] gene expression matrix: /home/andregabriel/Desktop/CIRI_test/gene_count_matrix.csv
[Wed 2021-03-17 22:11:31] [INFO ] Output DE results: /home/andregabriel/Desktop/CIRI_test/circRNA_de.tsv

Warning message:
In estimateDisp.default(y = y$counts, design = design, group = group,  :
  No residual df: setting dispersion to NA
Error in glmFit.default(y = y$counts, design = design, dispersion = dispersion,  : 
  Design matrix not of full rank.  The following coefficients not estimable:
 treatT
Calls: glmFit -> glmFit.DGEList -> glmFit -> glmFit.default
Execution halted
[Wed 2021-03-17 22:11:35] [INFO ] Finished!

[Kevinzjy]

Hi @andre-gabriel-42, could you show me the content of library_info.csv?

[andre-gabriel-42]

Hello @Kevinzjy Here is the content of my library_info.csv

Sample,Total,Mapped,Circular,Group,Subject
Case_A05,70994698,62755636,293486,T,A05
Case_A08,45394204,35150610,462928,T,A08
Case_A18,43519496,37203988,409090,T,A18
Ctrl_A09,41760562,33669376,845456,C,A09
Ctrl_A10,66158762,57916538,287584,C,A10
Ctrl_A11,33836070,26121594,512778,C,A11
Ctrl_A12,52129642,45188090,373526,C,A12
Ctrl_A13,34452232,25596474,442308,C,A13
Ctrl_A14,49450134,41324998,1001986,C,A14

[Kevinzjy]

Could you try removing the last column, and run CIRI_DE_replicate again?

[andre-gabriel-42]

It seems that I have to install statmod package

Command:

/home/andregabriel/.local/bin/CIRI_DE_replicate --lib library_info_without_last_column.csv --bsj circRNA_bsj.csv --gene transcript_count_matrix.csv --out circRNA_nolastcloumn.tsv

Log:

[Thu 2021-03-18 12:49:04] [INFO ] Library information: /home/andregabriel/Desktop/CIRI_test/library_info_without_last_column.csv
[Thu 2021-03-18 12:49:04] [INFO ] circRNA expression matrix: /home/andregabriel/Desktop/CIRI_test/circRNA_bsj.csv
[Thu 2021-03-18 12:49:04] [INFO ] gene expression matrix: /home/andregabriel/Desktop/CIRI_test/transcript_count_matrix.csv
[Thu 2021-03-18 12:49:04] [INFO ] Output DE results: /home/andregabriel/Desktop/CIRI_test/circRNA_nolastcloumn.tsv
Error in fitFDistRobustly(var, df1 = df, covariate = covariate, winsor.tail.p = winsor.tail.p) : 
  statmod package required but is not installed
Calls: estimateDisp ... estimateDisp.default -> squeezeVar -> fitFDistRobustly
Execution halted
[Thu 2021-03-18 12:49:25] [INFO ] Finished!

[Kevinzjy]

Yes, so just run install.packages("statmod") in R to install it :)

[andre-gabriel-42]

Success!! Thank you so much!

[Kevinzjy]

No worries.