Skip to main content
Topic: Automatic detecion of problems during analysis (Read 6036 times) previous topic - next topic

Automatic detecion of problems during analysis

Dear all,

I was wondering if anyone was aware of any software (or a proccess included in any software) that can automatically test for problems with your analysis during the analysis itself?
I image a tool that would check for drift in retention time and line broadining, continiously check the mass calibration, check for increases in common contaminants, dramatic drops in sensitivity etc and give a running report.
Does something like this exist? Or in part?


Any hints you can give me will will be highly appreciated. I am also interested to know if others would find such a tool useful.



- Jan Stanstrup.
Blog: stanstrup.github.io

Re: Automatic detecion of problems during analysis

Reply #1
it that something which integrates/checks while data acquired on an instrument? or something on the data side only, post acquisition, that you are thinking about. For the later part there are some softwares, depending on the technology that can have mean to asses the quality of the data.

Reza Salek

Re: Automatic detecion of problems during analysis

Reply #2
I was thinking of the first. A running analysis (updated when a new file is available) that would warn of potential problems so that you have a chance to catch problems when they appear and not 2 months later during data analysis. XCMS + shiny for a graphical report could do it was my thought.
Blog: stanstrup.github.io

 

Re: Automatic detecion of problems during analysis

Reply #3
Depending on which problems you are after to prevent (samples or QC/QA) and what is your instrument make-model and data type. Have a look at:
http://www.ncbi.nlm.nih.gov/pubmed/23088386
SIMPATIQCO: a server-based software suite which facilitates monitoring the time course of LC-MS performance metrics on Orbitrap instruments.
Again there are others  particularly if QC/QA is the target, which also can give feeling for sample quality or potential issues. They are mostly for proteomics data sets but should equally work for metabolomics data from similar instrument. 

Your idea for XCMC equally should work in a pipeline and with a mean of visualising and interrogating the results.

Reza