chromatogram refine issues

[g0079]

Hi all, I'm a beginner of XCMS. After learning XCMS, I'm going to try it out with a batch of my own data. When I get to the peak refine step, I get the following error message. My data was acquired as DDA on an Agilent 6545 QTOF. My R version is 4.2.1 and my XCMS version is 3.18.0. I tried to change the parameters of mnpp, only the error message changed, but the error is still reported. Could you please help me find out what the problem is? Thank you very much.

raw_file = fs::dir_ls("raw_data", recurse = T, glob = "*.mzML") sample_name = str_extract(basename(raw_file), pattern = ".*(?=\\.)") sample_group = c(rep("QC", 3), rep("OriginA", 4), rep("OriginB", 4), rep("OriginC", 4), rep("OriginD", 4)) pd = data.frame(sample_name, sample_group) raw_data = readMSData(raw_file, pdata = new("NAnnotatedDataFrame", pd), mode = "onDisk") cwp = CentWaveParam(ppm = 10, peakwidth = c(10, 40), noise = 1000) caasData = findChromPeaks(raw_data, param = cwp) mnpp = MergeNeighboringPeaksParam(expandRt = 4) caasData = refineChromPeaks(caasData, param = mnpp) Evaluating 702 peaks in file Sam3-2.mzML for merging ... Stop worker failed with the error: wrong args for environment subassignment Error: BiocParallel errors 0 remote errors, element index: 19 unevaluated and other errors first remote error: In addition: Warning messages: 1: In serialize(data, node$con) : 'package:stats' may not be available when loading 2: In serialize(data, node$con) : 'package:stats' may not be available when loading 3: In serialize(data, node$con) : 'package:stats' may not be available when loading 4: In serialize(data, node$con) : 'package:stats' may not be available when loading 5: In serialize(data, node$con) : 'package:stats' may not be available when loading

[JunYang2021]

I've just encountered this issue. I think you might be using R in windows, too. The reason is that R in Windows automatically open multiple processes parallelly to finish computation. But here one of R process may be closed or no longer accessible from the main thread. The solution is reducing the number of computation processes. You can use this code to get your physical cores number. parallel::detectCores(logical = FALSE) Then set processes number (less than total physical cores number): register(bpstart(SnowParam(2))) ## To start 2 processes for parallel processing(for example) Then you can try your refinement again. If it doesn't work, you can use register(SerialParam()) to cancel prarallel processing, but it would be slow.

[g0079]

I've just encountered this issue. I think you might be using R in windows, too. The reason is that R in Windows automatically open multiple processes parallelly to finish computation. But here one of R process may be closed or no longer accessible from the main thread. The solution is reducing the number of computation processes. You can use this code to get your physical cores number. parallel::detectCores(logical = FALSE) Then set processes number (less than total physical cores number): register(bpstart(SnowParam(2))) ## To start 2 processes for parallel processing(for example) Then you can try your refinement again. If it doesn't work, you can use register(SerialParam()) to cancel prarallel processing, but it would be slow.

It worked! In the end, I used register(SerialParam()) to process the data one by one, and no more errors were reported. I really can't think of this as the reason, can this be called a bug? Thank you very much!@JunYang2021

[JunYang2021]

I found this solution from this issue https://github.com/sneumann/xcms/issues/627. It was said that parallel processing in Windows is more easily to be corrupted. Maybe in Linux, this problem will not happen. @g0079

[jorainer]

Great! Thanks @JunYang2021 for answering and providing a solution!

I'm closing the issue now - feel free to re-open if needed.