Error with function cnvOncoPrint using a colorectal cancer dataset of 34 WES samples #34
Comments
|
Hi Efstathios,
Note that
This should work now using the new version of the CNVRanger package (version 1.11.2). You can install the new version 1.11.2 of the CNVRanger package in a fresh R session directly from github via BiocManager::install("waldronlab/CNVRanger") This works now on my end, when applying Hope that helps. Best, |
|
Dear @lgeistlinger, initially I would like to thank you for your response !! As I re-install the updated version, I repeated the analysis, but using this time also updated total copy numbers (as in my previous attached object I have included log2 segmented ratios) to improve inference; also for reproducibility, I have also attached the saved rds file of the updated input-using the same cohort:
Then, subset directly to the protein coding genes from above step: However, if you see from the above x.sel matrix, PPIAL4G gene is altered in 3 patients, whereas its frequency in the cgenes object from above, is 5; how you explained this discrepancy? Is any step from above wrong? Or there is another explanation for this difference? For example, the same applies for the second gene with symbol FAM72D; which in the final patient matrix is altered in 2 patients, whereas again showed a frequency of 5 samples in cgenes; Could this be attributed to the function populationRanges, where it summarizes individual calls to more "discrete CNV regions" that overlap between samples? And when I assign the frequency from the sel.grl.freq object, it is erroneous? And frequency should only be assessed at this stage? Or it should not be 5 in the above genes, but as in the final x.sel matrix? Just to pinpoint, in my input txt file, the only "deviations" from the tutorial were: instead chr1, chr2, the chromosome column value has 1, 2, etc.; In addition, I have include copy number states above 4, like 5, 6 as some samples have triploid phenotypes; Do I need anything from above 4 to convert it to 4, or it is not harmful? With Kind Regards, Efstathios |
Dear @lgeistlinger,
moving from email where I tried to reach you for a specific error appeared when trying to implement CNVRanger for 34 WES samples from colorectal cancer patients:
briefly, based on an international collaborative project between DKFZ and other Greek research institutions regarding personalized medicine in cancer (http://www.accc.gr/aboutACCC_info.html), we have acquired around 34 samples with DNA and RNA which were sequenced and pre-processed in the DKFZ bioinformatics facility. Overall, my major goal is to investigate, if there any specific mutational patterns in any of the 3 separated groups of patients, that might interrelate the presence of specific mutational patterns (KRASonly, BRAFonly and wild type), as the ultimate goal is to investigate the molecular landscape of these 3 defined groups-we also utilize additionally public TCGA data to enhance our sample size-
My major question would be for the actual analysis of the CNV data from our 34 patients-as already copy number alterations were called with cnvkit, and as a further step purity and ploidy estimated; for a start, I tried to analyze all data collectively, based on the segmented log2 rations, as in your vignette, with the following code:
Thus, in your opinion, how this error could be fixed? It is something wrong with any of the input files? For simplicity I have attached also the initial created GRangeslist object;
Input.test.grl.CRC.rds.zip
With Kind Regards,
Efstathios
The text was updated successfully, but these errors were encountered: