Re: Help required to fix an error in IDEOM

[Darren Creek]

Hi Sithara, 

Apologies for this bug. I think it arises during the normalisation step and it will hopefully be fixed in the next version. 

The problem is most likely due to an empty row at the bottom of the mzMatch data matrix. 

To avoid this problem:

Before running 'step 3'

 - select the 'mzMatch' sheet in IDEOM

 - scroll to the bottom and find the last row that has all '0' or missing values

 - delete that row using the right-click delete option

 - return to the 'settings' sheet and run step 3.

If this doesn't fix the problem then please double-check the mzMatch data matrix for other non-number values, and check the grouping information on the 'settings' sheet (columns H:L).

 Otherwise, if it still doesn't work please send me a copy of the offending file. (you can delete the data from the DB sheet to make it smaller for email). 

cheers,

Darren 

On Tue, Apr 2, 2013 at 7:09 AM, thaara pillai <thaara...@gmail.com> wrote:

Hi, Dr. Darren,

 

Thank you so much for developing the pipeline for the analysis of metabolomics data, it is really helpful and I am really enthusiastic to learn more about it.

 

I am a research trainee at CTSA metabolomics core at Mayo clinic, Minnesota. I am trying to analyze a highthroughput metabolomics data from ULPC QTOF. I have converted the.d files to mzXML file using MS convert from proteo wizard. Then used XCMS, mzMatch/ Peak ML viewer for further analysis. Then I tried to import the output text file from mzMatch to IDEOM v 18. I had an error while running the identification macro.

 

I am fresh graduate in MPH (Master of Public health) from India, have good knowledge in R, but I am very new to Metabolomics studies.

 

Could you please advice me a solution to fix this error. I have read your Ideom tutorial Part 1 & 2, also checked the discussion of similar error in Ideom google groups.

 

I have also tried to run the identification macro without changing the parameters, same error comes then. Please find below the commands from VBA, I couldn't understand the highlighted part.

 

 

Looking forward to hear from you soon

 

Thank you,

Sincerely,

Sithara

 

 

 For r = 1 To (groupcount)
    If (r + 2) <> blankrow And setsheet.Cells(2 + r, 8) <> "Exclude" And setsheet.Cells(2 + r, 8) <> "Standards" Then

            num = setsheet.Cells(2 + r, 11)
            posi = setsheet.Cells(2 + r, 12)
            Set samprange = adsheet.Range(Cells(q, 28), Cells(q, 27 + num)).Offset(0, posi)
            dgroupcount = Application.CountIf(samprange, ">0")

        If (Application.Average(samprange) > intensityfilter) And num > 1 Then     'avoids strict RSD filter applying to small peak groups and to avoid error if average = 0 or only 1 sample
            RSD = Application.StDev(samprange) / Application.Average(samprange)

            If (r + 2) = QCrow Then
                adsheet.Cells(q, 19) = Application.Round(RSD, 2)
         '       If RSD > maxRSD Then            ' check for maxRSD
             '       maxRSD = RSD
            End If

      '      Else
                If RSD > maxRSD Then            ' check for maxRSD
                    maxRSD = RSD
                End If

                If (r + 2) = QCcolrow Then          ' check for surrogate QC
                    adsheet.Cells(q, 19) = Application.Round(RSD, 2)
                End If
                                                    ' add 'Groups' if significant
                If (Application.min(samprange) > blankmax) And (RSD < RSDlimit) And dgroupcount = num Then
                        sigtext = sigtext & "'" & setsheet.Cells(2 + r, 10)
                End If
                                                    ' add #detections and maxintensity
                dcount = dgroupcount + dcount
                maxintensity = Application.max(samprange, maxintensity)
                                                    ' calculate control and treatment means
                If setsheet.Cells(2 + r, 8) = "Control" Then
                    contnum = num
                    Set controlrange = samprange
                    controlmean = Application.Average(controlrange)
                ElseIf setsheet.Cells(2 + r, 8) = "Treatment" Then
                    treatnum = num
                    Set treatrange = samprange
                    treatmean = Application.Average(treatrange)
                End If

         '   End If
                   
        Else                      ' to allow for set with single sample or low intensity to be sig
      '      If (r + 2) <> QCrow Then
                If Application.min(samprange) > blankmax Then
                    sigtext = sigtext & "'" & setsheet.Cells(2 + r, 10)
                End If
      '      End If
                                             ' add #detections and maxintensity
            dcount = dgroupcount + dcount
            maxintensity = Application.max(samprange, maxintensity)
                                                    ' calculate control and treatment means
            If setsheet.Cells(2 + r, 8) = "Control" Then
                    contnum = num
                    controlmean = Application.Average(samprange)
                    Set controlrange = samprange
            ElseIf setsheet.Cells(2 + r, 8) = "Treatment" Then
                    treatnum = num
                    treatmean = Application.Average(samprange)
                    Set treatrange = samprange
            End If
        End If
    End If
    Next r

'Groups'
adsheet.Cells(q, 12) = sigtext
'RSD
adsheet.Cells(q, 20) = Application.Round(maxRSD, 2)
'Detections
adsheet.Cells(q, 27) = dcount
'Maximum Intensity
If maxintensity = 0 Then
        adsheet.Cells(q, 21) = 0
    ElseIf maxintensity > 10 Then
        adsheet.Cells(q, 21) = Application.Round(maxintensity, 0)
    Else
        adsheet.Cells(q, 21) = maxintensity
End If

'Treat vs control
   If controlmean = 0 Then
    adsheet.Cells(q, 23) = treatmean / Minintensity
   Else
    adsheet.Cells(q, 23) = treatmean / controlmean
   End If

'Pvalue     'Welch's 2-tailed t-test (unequal variances assumed). default is unpaired but can be changed to paired on settings sheet
If (contnum > 1) And (treatnum > 1) Then
    If (controlmean = 0) And (treatmean = 0) Then
        adsheet.Cells(q, 24) = "ND"
    Else
        adsheet.Cells(q, 24) = Application.TTest(treatrange, controlrange, 2, statspair)
    End If
End If

'Fisher ratio (at end)
On Error Resume Next
If (contnum > 1) And (treatnum > 1) Then
    adsheet.Cells(q, adcolno) = Application.IfError((treatmean - controlmean) ^ 2 / ((Application.StDev(treatrange) ^ 2) + (Application.StDev(controlrange) ^ 2)), 0)
ElseIf (treatnum > 1) And (treatmean > controlmean * 2) Then
    adsheet.Cells(q, adcolno) = "high"
ElseIf (contnum > 1) And (controlmean > treatmean * 2) Then
    adsheet.Cells(q, adcolno) = "low"
End If
On Error GoTo 0
--

With regards,

 

Sithara Vivek,

Research trainee,

Metabolomics core,

Mayo clinic,

Rochester, MN - 55902

Mobile - 704 450 0343

Email - thaara...@gmail.com

               sit...@sctimst.ac.in

  

--
Dr Darren Creek

NHMRC Training Fellow
Department of Biochemistry and Molecular Biology, Bio21 Institute

University of Melbourne

dcr...@unimelb.edu.au

darre...@gmail.com

(+61) 3 83442351