elelab / cancer-biomarker-prediction-pipeline-campp Goto Github PK

retainedvar <- intersect(res.DE.names, rownames(sdata))

• does this make sense? Shouldn't it be:
  retainedvar <- intersect(res.DE.names,res.DEs.names)#DE names from second dataset.
  ?

Error in makeContrasts("arg.groupC3-TAG-arg.groupILC-Pik3Ca/Cdh1Null", :
The levels must by syntactically valid names in R, see help(make.names). Non-valid names: arg.groupC3-TAG,arg.groupILC-Pik3Ca/Cdh1Null,arg.groupKPB-1A,arg.groupKPB-1B,arg.groupKPB-Primary,arg.groupKPB25L-Apobec,arg.groupKPB25L-UV,arg.groupNML-BALBC,arg.groupNML-FVB,arg.groupp53-2151,arg.groupp53-2153F,arg.groupp53-2224,arg.groupp53-2224L,arg.groupp53-2225L,arg.groupp53-2245,arg.groupp53-2336R,arg.groupp53-9263-3,arg.groupp53-T1,arg.groupp53-T12,arg.groupp53-T2,arg.groupp53-T6,arg.groupp53-T7,arg.groupT11-Apobec,arg.groupT11-UV
Calls: eval -> eval -> eval -> makeContrasts
Execution halted

I have performed DAA on glycandata.xlsx and now want to run WGCNA on it. If I use glycandata.xlsx as the input(according to the CAMPP manual), it gives the following error.

Error: WGCNA may be performed with either results of differential abundance / expression analysis (DA / DE) or with all variables (ALL). N.B It is not advisable to run WGCNA with all variables if n > 5000. This will make WGCNA slow and plots will be difficult to iterpret. If 'ALL' is chosen and n > 5000, NO plots will be generated, but module variable interconnectivity scores will still be computed.
Execution halted

However on performing WGCNA on glycandata_DE.xlsx(result of DA Analysis), it says:

Warning messages:
1: In lapply(my.data, as.numeric) : NAs introduced by coercion
2: In lapply(my.data, as.numeric) : NAs introduced by coercion
3: In lapply(my.data, as.numeric) : NAs introduced by coercion

• The input data has between 37.5% - 37.5% missing values per row.

• Variables (rows) with more than 70% missing values will be removed.

• The input data has between 0% - 100% missing values per column.

• Samples (rows) with more than 80% missing values will be removed. Variables (rows) with more than 50% missing values are imputed using the overall mean per sample.
  ... Performing missing value imputation. N.B Uncertainty increases with number of missing values!.
  Error in ReplaceNAs(arg.data) : object 'varnames' not found
  Execution halted

What should the input for WGCN Analysis?

I want to obtain differentially expressed genes over samples of patients infected with a disease along with control samples. I have used the following command:
Rscript CAMPP.R -d data2.xlsx -m metadata2.xlsx -v array -g Accession,Code -t log2 -l 0.0 -s TRUE -f 0.6,0.05 -n DEG
However, this returns the following error:
Error: Aesthetics must be either length 1 or the same as the data (32): fill and colour
Backtrace:
█

1. └─ggplot2::ggsave(...)
2. ├─grid::grid.draw(plot)
3. └─ggplot2:::grid.draw.ggplot(plot)

4. ├─base::print(x)
   

5. └─ggplot2:::print.ggplot(x)
   

6.   ├─ggplot2::ggplot_build(x)
   

7.   └─ggplot2:::ggplot_build.ggplot(x)
   

8.     └─ggplot2:::by_layer(function(l, d) l$compute_aesthetics(d, plot))
   

9.       └─ggplot2:::f(l = layers[[i]], d = data[[i]])
   

10.         └─l$compute_aesthetics(d, plot)
    

11.           └─ggplot2:::f(..., self = self)
    

12.             └─ggplot2:::check_aesthetics(evaled, n)
    

Execution halted

Do I need some modifications in my data for this command to work?

I'm trying to run CAMPP.R "sudo Rscript CAMPP.R" and I get several dependency errors.

I haven't knowledge of the R language, I work on a linux platform.

cat /etc/issue
Ubuntu 18.04.2 LTS

R --version
R version 3.4.4 (2018-03-15) instalada por defecto mediante apt

sudo Rscript CAMPP.R &> >(grep -i error)
ERROR: dependency ‘locfit’ is not available for package ‘edgeR’
ERROR: dependency ‘car’ is not available for package ‘rstatix’
ERROR: dependency ‘mvtnorm’ is not available for package ‘maxstat’
ERROR: dependency ‘mvtnorm’ is not available for package ‘flexsurv’
ERROR: dependencies ‘cowplot’, ‘rstatix’ are not available for package ‘ggpubr’
ERROR: dependencies ‘ggpubr’, ‘maxstat’ are not available for package ‘survminer’
ERROR: dependency ‘car’ is not available for package ‘rstatix’
ERROR: dependency ‘mvtnorm’ is not available for package ‘maxstat’
ERROR: dependencies ‘cowplot’, ‘rstatix’ are not available for package ‘ggpubr’
ERROR: dependencies ‘ggpubr’, ‘maxstat’ are not available for package ‘survminer’
ERROR: dependency ‘latticeExtra’ is not available for package ‘Hmisc’
ERROR: dependency ‘Hmisc’ is not available for package ‘WGCNA’
ERROR: dependency ‘latticeExtra’ is not available for package ‘Hmisc’
ERROR: dependency ‘Hmisc’ is not available for package ‘WGCNA’
Error in FUN(X[[i]], ...) : there is no package called ‘edgeR’
Error: Ups! Something went wrong, one or more R-package dependencies are not installed correctly. Check the script CAMPPmissingpackages.R. Alternatively you can download and use the Packrat library freeze, see manual for specifics. Packrat library in brief:

Can CAMPP.R work on linux?

Does this make sense? For now, counts are intersected also on the sample level and this means zero intersect