Integration software?

[Peter_D]

Hi

I'm trying to find some simple software that will integrate the peaks in chromatogram files. I can convert the origiinal binary files to ascii format, and they are basically two columns, time and signal level. Example header of a file looks like this

0.00000 0.15781
0.00083 0.16055
0.00167 0.15898
0.00250 0.15977
0.00333 0.15977
0.00417 0.15977
0.00500 0.15977
0.00583 0.15977
0.00667 0.15937
0.00750 0.15898

This is just the first ten lines, each file has 7200 lines.

I'd like to find just the retention time, area and peak width of each peak in the chromatogram.

Example output table file should look like this

EXAMPLE TEXT OUTPUT FILE
ORIGINAL INJECTION TIME 19-FEB-2013 15:15
1,1.653000,636.400024,1.600000
2,3.623000,610.309998,1.450000
3,3.980000,22.600000,3.600000
4,4.107000,14.220000,1.350000
5,4.624000,708.869995,1.650000
6,5.546000,683.700012,1.750000
7,5.625000,1651.579956,1.800000
8,5.886000,755.000000,1.900000

The line format is:
peaknumber,retention time,peak area,peak width

I also need to process the files in large batches. Peak identification, calibrations etc. not required, just an accurate integration of the peaks.

Anyone have any ideas of software that I can use?

Thanks

Peter

[Peter Apps]

Why not use the software that the original data came from ?

I have heard that there are integration packages in both MATLAB and R number crunching software - I doubt that they qualify as "simple" though.

Peter

[Peter_D]

Why not use the software that the original data came from ?

I have heard that there are integration packages in both MATLAB and R number crunching software - I doubt that they qualify as "simple" though.

Peter

The original software doesn't produce individual text reports for each file, so I'll have to reprocess the data. The above example was integrated by hand, however I've got thousands of files to reprocess!

I've not heard of MATLAB but a colleague is looking for something in R.

I've looked at Openchrom, but it only does GC/MS data? Next version is supposed to have FID support which may be what I want.

Thanks

Peter

[shfo]

The original software doesn't produce individual text reports for each file, so I'll have to reprocess the data. The above example was integrated by hand, however I've got thousands of files to reprocess!

I've not heard of MATLAB but a colleague is looking for something in R.

I've looked at Openchrom, but it only does GC/MS data? Next version is supposed to have FID support which may be what I want.

Thanks

Peter

I haven't used any Matlab or R integration libraries, but I would guess that they require some (easy-ish) programming to get the results in the format you want. I would be interested in hearing about the open-source integration libraries out there, because I've found there's a lack of good example code out there to understand exactly how integration works.

I think the latest version of Openchrom should support Agilent FID files right now. I've been working on some more integration options in Aston (which also has support for FID files but not as many formats as Openchrom), but I haven't spent any time on reporting or batch-processing so you'd have the same problem as you do with your vendor's software. So I would say unless Openchrom can do the reports you want or you have some programming skills, reprocessing everything manually might be your best option (sorry!)

[gesen]

Try my software http://gesen.p.ht/ . It is possible to integrate with it, but you have to select peaks manually.

[Peter_D]

I haven't used any Matlab or R integration libraries, but I would guess that they require some (easy-ish) programming to get the results in the format you want. I would be interested in hearing about the open-source integration libraries out there, because I've found there's a lack of good example code out there to understand exactly how integration works.
I think the latest version of Openchrom should support Agilent FID files right now. I've been working on some more integration options in [url=aston.googlecode.com]Aston[/url] (which also has support for FID files but not as many formats as Openchrom), but I haven't spent any time on reporting or batch-processing so you'd have the same problem as you do with your vendor's software. So I would say unless Openchrom can do the reports you want or you have some programming skills, reprocessing everything manually might be your best option (sorry!)

Thanks for the reply, I've also been unable to find any good example code, even though the principles of chromatogram integration are well documented. Wish I knew how to write code.

I've been in touch with Openchrom, apparently version 0.8.0 will have better FID support and it should be able to run from the command line, with the files to be processed called from a ascii file. Won't be out for a while yet.

Pete

[Peter_D]

Try my software http://gesen.p.ht/ . It is possible to integrate with it, but you have to select peaks manually.

Thanks for the link, but I have thousands of chromatograms to reprocess. I need something that will will work without manual intervention on large batches, I can accept it wont be 100% all of the time, or if it falls over at some point I can reset any parameters and restart from there. Manual integration is out of the question. I have a collection of 8 chromatograms per hour over three years.

In the past I have used software on similar systems like Chrompack PCI (20 years ago!) or Perkin Elmer Totalchrom and they have worked quite well, but only with their own format of files. Just set up a reasonable set of integration parameters. It would be nice to have a program that could produce report like those.

Pete

[tom jupille]

How "clean" are your chromatograms? And how well resolved are your peaks? If you don't have a lot of noise or drift to deal with, and if all your peaks are baseline resolved, you could do it with VBA macros in Excel. The hard part is allocating areas (setting baselines) when peaks are partially resolved.

[Peter_D]

How "clean" are your chromatograms? And how well resolved are your peaks? If you don't have a lot of noise or drift to deal with, and if all your peaks are baseline resolved, you could do it with VBA macros in Excel. The hard part is allocating areas (setting baselines) when peaks are partially resolved.

Indeed, I can't guarantee that all the peaks are baseline resolved, all of the time. I would definitely need code capable of baseline/valley/skim etc. interpretation.

For noise, most software I've used has parameters adjustable for sensitivity and bunching/smoothing. The chromatograms I am looking at for this look more than reasonable for noise compared to files I have processed in the past.

I know chromatogram integration isn't simple mathematics, but i'm surprised there isn't some easily accessible code out here to do it.

Thanks

Peter

[CE Instruments]

Stupid question ? How did you acquire the data ? How did it end up in this format . If it was acquired by a proper data system would it not be easier to try and find someone who could re-process it for you ?

[Peter_D]

Stupid question ? How did you acquire the data ? How did it end up in this format . If it was acquired by a proper data system would it not be easier to try and find someone who could re-process it for you ?

Good question, it was acquired using software that saved the raw chromatograms as individual files but put any results into a large log file. The results generated are very poor and it's extremely tedious to reprocess any data using the original software.

I've got some other software that I can use to re-analyse the data, but it needs an invidual text report file for each chromatogram. That's why I need to find something that can accurately reintegrate the original data.

The analyser ran for three years generating 8 chromatograms an hour. So quite a lot of data to get through.

Thanks

Peter

[Klinge42]

Hi Peter,

until now OpenChrom does not support the processing of FID signals beyond viewing and simply converting the files. But this is planned for upcoming releases.
Have you found a solution yet?
If not I could try to figure out a work-around using OpenChrom and KNIME. Could you send me e.g. 5 ascii-files as examples, so I can see how complex the chromatograms are? You can contact me via the contacts page (andreas) on the OpenChrom-site.

Regards

[Peter_D]

Hi Andreas

I've already been in touch with Philip at OpenChrom.

I understand version 0.8.0 is due for release at the end of this month and will be able to process my FID .xy chromatograms.

Is this correct?

Regards

Peter