Error in Differential expression analysis
Hi, I'm experiencing an error while running the differential expression analysis. I downloaded miARma and installed it from source code, I have installed Bioconductor and all the packages (following the procedure in issue #55 and similar since I got the same error), and I was able tu run and finish the example of Hypoxia. So I replaced the example files with mine (fastq reads, genomes, annotations, indexes) and I run the analysis step by step. I was able to do the adapter removal, quality analysis, alignment with bowtie1 and readcount, but after running the DE with EdgeR or both EdgeR and NOIseq I got this error:
*Problem while running this R command:
print(resultsfiles)
Error: print(resultsfiles) : object 'resultsfiles' not found Execution halted*
I tried to reinstall bioconductor packages, to move files among directories, to see other similar issues but nothing worked for me, and I don't know why, since I correctly run the example and I cannot run my own analysis. Maybe I id some mistakes? I attached the log and stat files Please, help me, thanks
Alessandro
Comments (16)
-
-
Hi Eduardo, by trying your code I got this error:
Loading required package: edgeR Loading required package: limma Loading required package: ggplot2 Loading required package: gplots Attaching package: ‘gplots’ The following object is masked from ‘package:stats’: lowess Loading required package: genefilter Attaching package: ‘genefilter’ The following object is masked from ‘package:base’: anyNA 2017-10-16 09:13:27 Starting quality control analysis of TSA_cut_bw1 Error in scan(file, what, nmax, sep, dec, quote, skip, nlines, na.strings, : line 1 did not have 12 elements
Do you want me to attach contrast and targets files? They look correct to me. Alex
-
Hi, Please take a look first at the line1 of the ReadCount.tab file which is inside of the ReadCount_results folder. In this line you have to see the same sample than in the targets.txt file
If you see no error, please send the these 3 files: targets, contrast and the ReadCount.tab file
Thanks
-
Ok, in cut_bw1-ReadCount.tab file I see that I have 12 samples (4 samples that are mine + the 8 of the example) while the targets and contrast files have just 4 samples, so I think this is the origin of the error. I'm trying to fix the problem, run the analysis again and I'll let you know. Thank you
-
Change the output_dir and re-run. This should be enough
-
- attached targets.txt
- attached cut_bw1-ReadCount.tab
- attached contrast.txt
- attached nat_bw1-ReadCount.tab
-
Still same error. Do you think there's a problem in read count? This file seems to me very poor... Furthermore, running thr R procedure you typed in your first answer, now I got a different error:
2017-10-16 11:25:47 Starting quality control analysis of TSA_cut_bw1 Error in plot(density(log10(data[, 1])), col = boxcol[1], main = paste("Density plot of number of reads of ", : error in evaluating the argument 'x' in selecting a method for function 'plot': Error in density.default(log10(data[, 1])) : need at least 2 points to select a bandwidth automatically
-
ummm it seems that you have a low nombre of reads mapped or counted. two possibilities: 1. The strand. May be you use a wrong strand (this affects a lot the number of counted reads). You can change the stand in the ini file and rerun just the ReadCount process 2. Low mapped. Inside the *_readcount_results you will have some .summary files (one per sample). In this file you have a summary of counted and uncounted reads
Could you include the summary_xxx.xls file ? That is a summary of your results Let me know
-
- attached summary_results_TSA_miARma.xls
Done. Having a quick look, I saw that it includes the example files analyses as well...
-
Yes, unless you have use a fresh output_dir, all processed samples will be included.
Anyway, your % seems fine, why did you say that the file was poor ?
-
It was just a perception, cause I got used to tables full of data. Should it be due to the small number of replicates (2 for control and 2 for treated)??? I am just trying to reproduce the experiment taken from literature, so I am not understanding if it is a "reproducibility" problem or if I am not good in performing the analysis...
-
As you said, 2 is not enough for a good analysis, but is should be enough for getting the results.
But you should take a look to all details in the "real" experiment :
- filtering
- normalization method
- quality filtering
- usage of multimaping and overlapping reads
-
Hi Eduardo, I think I finally got the error. As I was telling you, my Readcount table looked to me poor. In fact, I got an error in drawing the density plot, since the program couldn't create a density plot with just 1 number (feature). So, I think I am missing some features in the read count table, since I got just one row/feature for my 4 samples. I think the problem came when I created the index with Bowtie (I didn't set a correct download). I'm gonna try it again and I will let you know if this was the real original problem.
-
Perfect !
-
Any update Alessandro ? Could I close the issue ?
-
- changed status to closed
No answer
- Log in to comment
Hi, Is the stat file empty?
Try this from a terminal:
You have to see the error. Please take a look to the target/contrast file, most of the errors come from those files Edu