Why would you use Process Only Sample Sets?

Empower2018

From my understanding, you can combine different injections and channels etc and create a process only sample set where you can alter details like batch number and add functions but cant change stuff like injection volume and method set etc. So how does that work if you process a process only sample set from injections taken from two different sample sets which have different method sets/instruments associated with them? Or can you only select the channels you want across the one sample set?

I cant see why you would want to do this, but someone must be using it. 

shaunwat

Hi Empower2018: I have not attempted this across different projects personally, but I don't see any reason why it would not function the same. You would need to select the projects you want to work with from the project tree shown in configuration so that you can pull data from multiple projects into one. But other than that...

I have used process only sample sets a few other times in the past when I was working on custom field development and realized that samples needed to be run in a specific order so that custom fields would calculate properly. However. after I got those methods setup in Empower, we also quickly revised the test methods to state that samples must be run in a specific order so as to avoid the use of processing only sample sets.

I would think that as an exception an auditor would be okay with using processing only sample set. However, if it became a habit I would think an auditor would start to wonder if the analysts were dropping injections that they didn't like, or otherwise piecing failing runs together so that they pass. Of course <711> and <905> would be another exception here with the stage testing in terms of being both habitual and okay to do.

shaunwat

One such reason to do this might be dissolution testing as an example.

The analyst goes to setup dissolution stage one testing and generates a failure. They then must move into stage two and perform calculations both on stage one and two results.

Say at stage two the analyst has another failure, at which point they must move onto stage three and calculate data using both stage one and two results.

At stage three lets say they pass and can finally move on...

Another example might be uniformity of dosage units. Again, an analyst would setup for stage one (n=10) and say a failure is generated. The analyst would then move on to stage 2 (n=20), but calculate results based upon stage one and two data (n=30).

USP <711> and <905> are prefect reasons for processing only sample sets.

If you have a mix up in channels or processing methods, Empower won't necessarily catch it; Empower just won't process the data it can't match up and leave you scratching your head as to why some results didn't generate.

When sample sets are processes, Empower ignores any instrument information. Though you do have the option to process by method set, you may also process to processing method only. Even when processing by MS, Empower only looks at the PM and derived channels as that is the only information required to process.

Empower2018

Hi shaunwat thanks for the reply? Have you used this function much? 

I can see how it might come in handy for the occasions you mention, ill keep that in mind for when I might need it in our lab. I wonder have you tried to do it across different projects? I presume you can select different channels across different projects and as you said, if it ignores any instrument lines then any processing method will do as long as it makes sense to your samples. 

MikeFink

Another useful purpose is for imported CDF data. CDF data is imported as individual channels and in order to process as a batch you need to create a process only sample set. 

Empower2018

Thanks shaunwat and others who replied, its a lot clearer now.