chasemc/IDBacApp

IDBac text file analysis

Closed this issue · 8 comments

twincacca commented

Hello,

we tried processing text files containing two columns of masses and intensities (separated by space/tabs, no header), by using the option "Click here to convert txt files".

IDBac detects the files (fig. 2.png) in the target directory and appears to process them successfully (fig. 3.png).

However, when trying to retrieve the analyzed data, only one sample appears in the results table (fig. 4.png, only 1 sample appears in the table ("LMA_Varia_") instead of the 6 supposed to be analyzed).

The procedure has been tested on a Windows machine via the installation with "IDBac.Setup.1.1.10.exe", as well as with the R installation on a Linux machine (ubuntu 16.04), both ways leading to an identical outcome.

We suspect this issues is related to the use of txt files instead of mzXM/mzXML files.

Please let us know if this is an issue of IDBac or else.

Thanks.

Cheers,
Antoine

1
2
3
4

chasemc commented

Thanks for the thorough details, I'm trying to get an update out be the end of the week and will try to include a fix for this (might be related to #128).

If possible, can you zip the files and attach drag and drop them into a reply here? Or can contact and send them privately through slack: https://join.slack.com/t/idbacapp/shared_invite/enQtNjkwNzYzNzA0MjE1LWI3Y2E3NGNlMjAzZDU4MTc3YjY4YmI3YzUwNWZjOGZjYzhlMmNkZDE3ZDY4Y2RlYTJhOWVlYjM4YzEwZTBjYzg

twincacca commented
chasemc commented

Thanks! Looks like you replied via email and GitHub didn't accept the attachment. Can you place it in a reply from GitHub.com? Link: #127

chasemc commented

I don't foresee making the fix this week as things got busy. However it looks like your data is centroid and IDBac requires profile (full-spectrum) data.

As a temporary solution if you are semi-comfortable with R and have profile data in .txt, .csv, or .tab you can:

Point inputDirectory to where your files are
Point ouputDirectory to where mzml files should be created

inputDirectory <- "F:/Chase/IDBac_Dev/example_data/tab/p"
ouputDirectory <- "F:/Chase/IDBac_Dev/abc"

Install MALDIquantForeign if you haven't

install.packages("MALDIquantForeign")

Then run:

source("https://gist.githubusercontent.com/chasemc/9391c733b8412c12ba0bc616a4edc275/raw/d0fbbf9de9202ec3182d135cb558a39154c3629a/convert_text_to_mzml.R")

converter(inputDirectory,
          ouputDirectory)

The resulting mzML files can then be used for IDBac

twincacca commented

Thanks for the suggestion, I could easily convert my data files from the format "txt" to "mzML" by following your instructions.

However, when I try to run IDBac using the mzML files, the same thing happens as when I previously run it with txt files. IDBac analyses only the first of the samples listed in the data containing folder.

Here I attach the mzXM data corresponding to the txt data I attached in my previous message. If you can have a look once you have time, it would be great.

Thanks a lot.

Cheers
Antoine
test_antoine_20190713.zip

chasemc commented

It looks like the data is still centroided (a list of peaks and intensities) which isn't currently supported. Do have the full-spectrum data for these samples?

twincacca commented

OK, I guess it's clear now. So, we need to work with full-spectrum data then. Yes, no problem. Thank you.

  • 👍1