workflow4metabolomics / metams Goto Github PK

MSP file when uploaded from user computer by Galaxy, are concidered as txt file and that format is not declared in the user_defined option. Please change that according to what was done in GC_default

When you want to run metaMS_runGC against a DB but you search also unknowns and that only 1 unknown was found you obtain an error :

Error in x[, settings$timeComparison] : incorrect number of dimensions
Calls: runGC ... match.unannot.patterns -> sapply -> sapply -> lapply -> FUN -> mean

This time this is not due to a variable which is a matrix whereas it should be a list but it apparently due to a list with only 1 thing in it... So like this, we have something like :

$`./0205065.CDF`
        mz      maxo      rt
CP0113  84 0.2200641 26.6531
CP0742 267 0.2948930 26.6531
CP0760 282 0.2927542 26.6656
CP0761 285 0.4139154 26.6656
CP0776 300 0.7185312 26.6656
CP0803 342 0.3021456 26.6656

$`./0205065_1.CDF`
        mz      maxo      rt
CP0920  84 0.2200641 26.6531
CP1549 267 0.2948930 26.6531
CP1567 282 0.2927542 26.6656
CP1568 285 0.4139154 26.6656
CP1583 300 0.7185312 26.6656
CP1610 342 0.3021456 26.6656

whereas we should have something like :

$`./0205004.CDF`
$`./0205004.CDF`[[1]]
        mz       maxo       rt
CP0004  51 0.01793574 24.51583
CP0012  52 0.01412946 24.51583
CP0016  53 0.04252074 24.51583
CP0024  54 0.01917578 24.52208
CP0031  55 0.07008467 24.50958
CP0035  56 0.01895593 24.50958
CP0044  57 0.02532629 24.50958

$`./0205004.CDF`[[2]]
        mz      maxo       rt
CP0365 121 0.9064138 24.74080
CP0933 236 0.1357717 24.74080
CP1076 286 0.2732609 24.74080
CP1233 359 0.1131362 24.74080
CP1350 415 0.1342479 24.74703
CP1461 470 0.1736152 24.74080
CP1479 478 0.1189479 24.74703
CP1504 496 0.1009723 24.74703

$`./0205004.CDF`[[3]]
        mz      maxo       rt
CP0300 109 0.4692890 24.49708
CP0427 133 0.6552011 24.49083
CP0702 183 0.4586975 24.49708
CP0714 184 0.2844711 24.49708
CP0766 194 0.1315781 24.49708
CP0979 249 0.1245335 24.49708
CP1179 329 0.1040561 24.49083
CP1285 385 0.1245118 24.48458

With the name of file first, then the different pseudospectra.

When files haven't any results between them, we have an error when we try to run sweep function :

Error in FUN(x, aperm(array(STATS, dims[perm]), order(perm)), ...) : 
  non-numeric argument to binary operator
Calls: runGC -> sweep```

The file input for csv, txt, msp and other file sneeded by metaMS is not efficient and cause serious issues.
TO DO add tests for each read.tables to check for the right format (sep ="", or ";" or"\t" ...

When the user specify a MSP library as input the names of the target compounds found are not written in EICs, only unknown X is specified anf it should be compound name Y for file X

Currently, RI files are detected as gg datatype (galaxy.datatypes.genetics:GenomeGraphs)
This format is not related to RI file but we added it to ease the tool usage.

Due to an update of mzR, mzData files or now not supported... => sneumann/mzR#253

Should change the test data and update the script to not able users to add it

Should be good to add this option in xml file for runGC

For now the links to How To pdf are targeted to a pydio based in Roscoff what is the new path in Toulouse?

with files names like XXX R&T - 1.cdf I had an issue with dataMatrix.tsv only containing variable names and no information on area. Those information stayed in variableMetadata.tsv due to some kind of issue during file formating (copy/paste columns..)

When you run it on one file, you obtain the following error :
Error in xset.l[[1]] : this S4 class is not subsettable

after normalisation Unknown X became Unknown.X and this is a problem
can we manage to avoid the checkname or in case of such small difference allow the tool to run

We have some blank outputs for EICs on the new usegalaxy instance

@yguitton @jsaintvanne

one error when we try to compare to house DB with RI filter active :

error when using Use RI as filter optioWarning message:
replacing previous import 'xcms::plot' by 'graphics::plot' when loading 'CAMERA'
Note: you might want to set/adjust the 'sampclass' of the returned xcmSet object before proceeding with the analysis.
< -------- Experiment of 42 samples ------------------------------------ >
< -------- Instrument: GALAXY.GC --------------------------------------- >
< -------- Annotation using database of 1 spectra ---------------------- >
< -------- Using xcmsSet object - only doing annotation ---------------- >
< -------- Removing artefacts () --------------------------------------- >
< -------- Matching with database of standards ------------------------- >
Error in h(simpleError(msg, call)) :
error in evaluating the argument 'x' in selecting a method for function 'which': non-numeric argument to binary operator
Calls: runGC ... lapply -> FUN -> which -> outer -> .handleSimpleError -> h
Execution haltedn

Hi,

When we use CDF files, it is not possible to obtain automatically their polarity... It's a problem during generations of TIC, BPC and EIC...

Actually the output table of unkn is a fixed table which is ranked by metaMS with no real order.
It can be a solution to find easily some compounds to have an interactive table in html format (like GOLM tool in W4M Galaxy) where users can sort by RT or MZ or something else to find their compound.

This issue merge with this one can be able users to find IDs of their compounds and then run EICs creations with these ones.

@yguitton @lecorguille

We obtain an error when DB and files don't really match... It is probably due to something resolve in this PR : yguitton/metaMS#14

But something stay wrong sometimes after... Not really the same number of rows between each column when we constructExpPseudoSpectra

Have to be explored !

add '$myparameter' for all script input

Hi,

A "small" bug in the wrapper seems responsible for the following issue, dataMatrix is only filled with variableNames.
I think that make.names is responsible for that, or maybe the file extension .mzML

@jsaintvanne, any idea?
Yann

During the runGC with CDF files and comparing with a DB, we obtain this error :

Error in x[, settings$timeComparison] : incorrect number of dimensions
Calls: runGC ... match.unannot.patterns -> sapply -> sapply -> lapply -> FUN -> mean

https://github.com/rwehrens/metaMS/blob/master/R/matchSamples2Samples.R#L173
It looks that due to a list of 1 element that is not a list but a matrix (probably because there is only one element in it)...
So I should modify how the list has been made just here https://github.com/rwehrens/metaMS/blob/master/R/matchSamples2Samples.R#L31

When using the RI option with a tab separated file or comma the runGC will crash saying that it need a matrix not a data.frame.
Idea modification of the RIarg<-read.table line to create a matrix with first column= rt and second column RI

some msp files are not fully compatible with metaMS
for exemple msp file fom AMDIS have ( mz int) instead of mz int; as mass spectrum descriptor

the idea for wrapper is 1 load msp file with a new read.msp that can deal with more msp format
then use write.msp format to create a converted msp file

To be able to find ID of unkn when you want to create EICs, it can be a good solution to add a column with them when we output the table.

@yguitton @lecorguille

TODO : Add test on EICs (problem with gs not found whereas I added pip install ghostscript and bgs in .travis.yml)

Currently, msp files are detected as txt files.
It may be interresting to create a specific sniffer.

For metaMS ploting function we need access to raw files and when files come from the xcmsSet.merge process the filepath doesn't includes the galaxy path

from metaMS (or zip)
resGC$xset[[c]]@xcmsSet@filepaths
[1] "/work/project/w4m/galaxy4metabolomics/galaxy-dist/database/jobs_directory/000/107/107131/working/./FWS_100perNaCl/alg10.cdf"

from merge

xset@filepaths
[1] "./alg3.mzData" "./alg2.mzData" "./alg11.mzData" "./alg9.mzData"
[5] "./alg8.mzData" "./alg7.mzData"

metaMS in galaxy do not allow user to define all the parameters and too many stays as default.
TO DO increase the available parameters

the actual 2.1.1 version on production now is not in github...sorrry for that

Today the batch Golm Data base search is not fully taking advantage of the RI filter available on Golm in fact we should trick the tool by setting a default RI = 1500 and add a shift of +/-2500.
can we add a function that will read for each compound in the *.msp file its own std.RI and use it as reference RI?

metaMS in R allows the use of RI values instead of rt for peaks comparison/peakmatching, But the optionis not activated in metaMs

In R metaMS the option remove Irrelevant peaks exists but is not currently functionnal in metaMS w4M

IDEA can we create a tool that read the Golm Metabolom database files and format it to be usable into metaMS ?

sometime the GCMS EIC.pdf generated by plotUnknown function is empty because the function crash, can we add in xml a boolean saying plotEIC or Not AND add in metams.r a tryCatch(plotUnknown) and in case of crash generate an empty pdf

We had a question from user concerning a strange error in metaMS with default parameters it crash (pdf empty) but not with user defined. the only difference is simthresh 0.95 in GC default and 0.7 in user defined.... any idea

When the first file has no match in DB, we obtain this table

$`./0205065.CDF`
[1] pattern      alternatives
<0 rows> (or 0-length row.names)

But we would like something with annotation column like this one :

$`./0205003.CDF`
  pattern annotation alternatives
1       1         -1             
2       3         -2             
3       4         -3             
4       5         -4             
5       7         -5             

Because here we bind first annotation and then new.annotation (this one contain new column) :
https://github.com/rwehrens/metaMS/blob/master/R/matchSamples2Samples.R#L160

Just changing the first binding table can be able to add this column... but does it work good with it ?
Testing that..