search

Nesvilab / FragPipe

A cross-platform proteomics data analysis suite
http://fragpipe.nesvilab.org
Other
195 stars 38 forks source link

Experimental Grouping for separate normalization #258

Open 1-stephan opened 3 years ago

1-stephan commented 3 years ago

Hi,

I want to analyze a project which includes 2 fractions, membrane and soluble proteins from the same samples.

For this project, LFQ normalization should be performed separately for the membrane and soluble fractions. Therefore, I would like to know if there is an option for this in fragpipe or not. It is not planned to directly compare soluble and membrane fractions. I will have two comparisons membrane KO vs WT and soluble KO vs WT.

In Maxquant, I would define 2 experimental groups in this case and I use separate LFQ normalization for each group. This has the advantage that I have evrything in one table instead of 2 separate tables. Additionally MBR is still used between the fractions which also boost the ID numbers a bit.

I really like fragpipe and results even look slightly better than Maxquant.

Thanks for your help, Stephan

fcyu commented 3 years ago

Hi Stephan,

Thanks for your interest in our tools. Unfortunately, we don't have this functionality. You may uncheck normalization for now.

Best,

Fengchao

anesvi commented 3 years ago

Stephan, You can control MBR, and allow MBR between runs from two experiments only with setting ‘top MBR runs’ to 1 ( in your case, since you only have 2 fractions in each). I think it should be a good way. In addition, You may want to use top-N instead of maxLFQ option

Sent from my iPhone

On Nov 3, 2020, at 8:19 AM, Fengchao notifications@github.com wrote:

 External Email - Use Caution

Hi Stephan,

Thanks for your interest in our tools. Unfortunately, we don't have this functionality. You may uncheck normalization for now.

Best,

Fengchao

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/Nesvilab/FragPipe/issues/258#issuecomment-721110604, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AIIMM6YMZR3DUDQD32TYQILSN77NNANCNFSM4TISH4AQ.

Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues

1-stephan commented 3 years ago

Thanks a lot for the comments. I will check the options for the quantification settings.

Especially for subcellular fractionation, I think many users would appreciate a grouping functionality and the option for separate normalization within in the individual groups. When you have one output table with all samples of different fractions, it is more convinient than separate tables for each fraction.

However, I don't know how complicated this would be to include in fragpipe.

Anyway, I am very happy with your tool. Keep going!

Thanks, Stephan

anesvi commented 3 years ago

Hi Stephan

What are other scenarios where such an option of defining parameter groups and doing LFQ separately for each set is useful, in your experience?

But we will discuss with Fengchao what we can do.

Thanks Alexey

Sent from my iPhone

On Nov 3, 2020, at 9:01 AM, 1-stephan notifications@github.com wrote:

 External Email - Use Caution

Thanks a lot for the comments. I will check the options for the quantification settings.

Especially for subcellular fractionation, I think many users would appreciate a grouping functionality and the option for separate normalization within in the individual groups. When you have one output table with all samples of different fractions, it is more convinient than separate tables for each fraction.

However, I don't know how complicated this would be to include in fragpipe.

Anyway, I am very happy with your tool. Keep going!

Thanks, Stephan

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://github.com/Nesvilab/FragPipe/issues/258#issuecomment-721132670, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AIIMM62WSB5KQEIMJ4U42QDSOAEHBANCNFSM4TISH4AQ.

Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues

1-stephan commented 3 years ago

Hi Alexey,

basically it is always useful when you have different sample types that belong to the same experiment. This might be:

For our lab, it is the first 2 scenarios, because we do a lot of secretome analysis.

From Maxquant users and their publications, I know that some people like to use this option to take old experiments as a basis for spectrum matching (i.e. when running Boxcar or if you already have a dataset with lots of fractions).

I don't know if also other users would like to have this function. I guess it depends on the effort to implement it. I would say, it is a "nice to have" option, but not a "must have".

I appreciate that you directly pick up new suggestions.

Thanks Stephan