search

Nesvilab / FragPipe

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

LFQ-MBR get less non-zero-intensity proteins with MBR ion FDR < 0.01 filtering compared to MaxQuant #449

Closed huangyingzhang closed 3 years ago

huangyingzhang commented 3 years ago

Hi,

I am using the FragPipe LFQ-MBR workflow to do LFQ as tutorial (https://fragpipe.nesvilab.org/docs/tutorial_lfq.html). The problem is that FragPipe identified fewer proteins comparing with Maxquant Match between runs(eg. ~1000 by FragPipe vs 2000 by Maxquant).
I don't if I need to optimize the parameters of the LFQ-MBR workflow or not and how to optimize. Would you please help me to check the workflow I used in the attached to see if there are some parameters that I can optimize? Thanks a lot!

Regards, Yingzhang

fcyu commented 3 years ago

Can you also send us your mqpar.xml file? Since you are doing a company, we need to make sure that the parameters are as close as possible.

Thanks,

Fengchao

On Tue, 31 Aug 2021 at 7:37 AM, huangyingzhang @.***> wrote:

  • Describe the issue or question:

Hi,

I am using the FragPipe LFQ-MBR workflow to do LFQ as tutorial ( https://fragpipe.nesvilab.org/docs/tutorial_lfq.html). The problem is that FragPipe identified fewer proteins comparing with Maxquant Match between runs(eg. ~1000 by FragPipe vs 2000 by Maxquant). I don't if I need to optimize the parameters of the LFQ-MBR workflow or not and how to optimize. Would you please help me to check the workflow I used in the attached to see if there are some parameters that I can optimize? Thanks a lot!

Regards, Yingzhang

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/449, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABU27W6WOLUKLE3PZZKEI4TT7S5HBANCNFSM5DD5GK5A . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

-- Dr. Fengchao Yu Research Investigator University of Michigan

huangyingzhang commented 3 years ago

Dear Dr. Yu,

Please check the attached mqpar.xml. And please be noted that the additional two fasta files (ARA-PEPs_peptides.fasta and ATSO.non_coding.pep.nr.cd-hit_aS1_aL0.3_c1.out.nr.fasta) which did not be used in FragPipe only contribute less than 20 proteins and are used to identify sORFs. Thanks a lot!

Regards, Yingzhang

On 31 Aug 2021, at 7:48 PM, Fengchao @.**@.>> wrote:

Can you also send us your mqpar.xml file? Since you are doing a company, we need to make sure that the parameters are as close as possible.

Thanks,

Fengchao

On Tue, 31 Aug 2021 at 7:37 AM, huangyingzhang @.***> wrote:

  • Describe the issue or question:

Hi,

I am using the FragPipe LFQ-MBR workflow to do LFQ as tutorial ( https://fragpipe.nesvilab.org/docs/tutorial_lfq.html). The problem is that FragPipe identified fewer proteins comparing with Maxquant Match between runs(eg. ~1000 by FragPipe vs 2000 by Maxquant). I don't if I need to optimize the parameters of the LFQ-MBR workflow or not and how to optimize. Would you please help me to check the workflow I used in the attached to see if there are some parameters that I can optimize? Thanks a lot!

Regards, Yingzhang

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/449, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABU27W6WOLUKLE3PZZKEI4TT7S5HBANCNFSM5DD5GK5A . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

-- Dr. Fengchao Yu Research Investigator University of Michigan

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/Nesvilab/FragPipe/issues/449#issuecomment-909162436, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AKLH75OMHXKB5BRJJNO3WYTT7S6SJANCNFSM5DD5GK5A. Triage notifications on the go with GitHub Mobile for iOShttps://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Androidhttps://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

fcyu commented 3 years ago

There is no file attached.

huangyingzhang commented 3 years ago

Dear Dr. Yu, Sorry about the mistake. Thanks, mqpar.txt

Yingzhang

fcyu commented 3 years ago

Hi Yingzhang,

Thanks for your mqpar file. Two tools' parameters look similar except for the variable modifications. FragPipe has oxidation and acetylation, but MaxQuant has oxidation, acetylation, deamidation, and Gln->pyro-Glu. I am not sure if the last two modifications matter much.

If you want us to take a closer look, you can send all of your files, including raw files, result files, and log files, to us. I can give you a link to upload if you want.

Best,

Fengchao

huangyingzhang commented 3 years ago

Hi Dr Yu,

Thank you very much.It would be very appreciate if you can take a closer look.

Best, Yingzhang

On 31 Aug 2021, at 9:35 PM, Fengchao @.**@.>> wrote:

Hi Yingzhang,

Thanks for your mqpar file. Two tools' parameters look similar except for the variable modifications. FragPipe has oxidation and acetylation, but MaxQuant has oxidation, acetylation, deamidation, and Gln->pyro-Glu. I am not sure if the last two modifications matter much.

If you want us to take a closer look, you can send all of your files, including raw files, result files, and log files, to us. I can give you a link to upload if you want.

Best,

Fengchao

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/Nesvilab/FragPipe/issues/449#issuecomment-909244059, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AKLH75JR42LBVG2X5PQVPQTT7TLCPANCNFSM5DD5GK5A. Triage notifications on the go with GitHub Mobile for iOShttps://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Androidhttps://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

fcyu commented 3 years ago

Sounds good. Please upload the files to https://www.dropbox.com/request/4qYaxBoV5vz6pu0HTWmw

Thanks,

Fengchao

On Tue, Aug 31, 2021 at 8:48 PM huangyingzhang @.***> wrote:

Hi Dr Yu,

Thank you very much.It would be very appreciate if you can take a closer look.

Best, Yingzhang

On 31 Aug 2021, at 9:35 PM, Fengchao @.**@.>> wrote:

Hi Yingzhang,

Thanks for your mqpar file. Two tools' parameters look similar except for the variable modifications. FragPipe has oxidation and acetylation, but MaxQuant has oxidation, acetylation, deamidation, and Gln->pyro-Glu. I am not sure if the last two modifications matter much.

If you want us to take a closer look, you can send all of your files, including raw files, result files, and log files, to us. I can give you a link to upload if you want.

Best,

Fengchao

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub< https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-909244059>, or unsubscribe< https://github.com/notifications/unsubscribe-auth/AKLH75JR42LBVG2X5PQVPQTT7TLCPANCNFSM5DD5GK5A>.

Triage notifications on the go with GitHub Mobile for iOS< https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android< https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-909766709, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABU27WYKNB5VKHXJ3WBE5LDT7VZ45ANCNFSM5DD5GK5A . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

-- Dr. Fengchao Yu Research Investigator University of Michigan

huangyingzhang commented 3 years ago

Hi, Dr. Yu,

I have uploaded all data. Thanks a lot!

Yingzhang

On 1 Sep 2021, at 9:01 AM, Fengchao @.**@.>> wrote:

Sounds good. Please upload the files to https://www.dropbox.com/request/4qYaxBoV5vz6pu0HTWmw

Thanks,

Fengchao

On Tue, Aug 31, 2021 at 8:48 PM huangyingzhang @.***> wrote:

Hi Dr Yu,

Thank you very much.It would be very appreciate if you can take a closer look.

Best, Yingzhang

On 31 Aug 2021, at 9:35 PM, Fengchao @.**@.>> wrote:

Hi Yingzhang,

Thanks for your mqpar file. Two tools' parameters look similar except for the variable modifications. FragPipe has oxidation and acetylation, but MaxQuant has oxidation, acetylation, deamidation, and Gln->pyro-Glu. I am not sure if the last two modifications matter much.

If you want us to take a closer look, you can send all of your files, including raw files, result files, and log files, to us. I can give you a link to upload if you want.

Best,

Fengchao

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub< https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-909244059>, or unsubscribe< https://github.com/notifications/unsubscribe-auth/AKLH75JR42LBVG2X5PQVPQTT7TLCPANCNFSM5DD5GK5A>.

Triage notifications on the go with GitHub Mobile for iOS< https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android< https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-909766709, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABU27WYKNB5VKHXJ3WBE5LDT7VZ45ANCNFSM5DD5GK5A . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

-- Dr. Fengchao Yu Research Investigator University of Michigan

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/Nesvilab/FragPipe/issues/449#issuecomment-909773381, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AKLH75KJ7AAIAM6CKNBSF2LT7V3ONANCNFSM5DD5GK5A. Triage notifications on the go with GitHub Mobile for iOShttps://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Androidhttps://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

fcyu commented 3 years ago

Thanks for your file.

In your proteinGroups.txt from MaxQuant, there are 11605 protein groups after filtering out "Only identified by site", "Reverse", and "Potential contaminant".

In your combined_protein.tsv from FragPipe, there are 13020 proteins.

To me, FragPipe found more protein than MaxQuant. Can you double check and elaborate more about your issue?

Best,

Fengchao

huangyingzhang commented 3 years ago

Hi, Dr. Yu,

Yes. There are more protein groups from FragPipe. But what I am focus on is the protein group identified in individual sample other than total proteins. For example, There is 2348 proteins with unique peptides in sample named G-1 from Maxquant proteinGroups result, while the same sample in FragPipe is 478 proteins with Unique Spectral count and 200 proteins with Unique intensity.

Regards, Yingzhang

On 1 Sep 2021, at 11:16 PM, Fengchao @.**@.>> wrote:

Thanks for your file.

In your proteinGroups.txt from MaxQuant, there are 11605 protein groups after filtering out "Only identified by site", "Reverse", and "Potential contaminant".

In your combined_protein.tsv from FragPipe, there are 13020 proteins.

To me, FragPipe found more protein than MaxQuant. Can you double check and elaborate more about your issue?

Best,

Fengchao

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/Nesvilab/FragPipe/issues/449#issuecomment-910385104, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AKLH75IQ3VVL63ORBEGDZE3T7Y7TZANCNFSM5DD5GK5A. Triage notifications on the go with GitHub Mobile for iOShttps://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Androidhttps://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

fcyu commented 3 years ago

Hi Yingzhang,

First of all, the unique peptides in MaxQuant include those from MS/MS and MBR, while the unique spectral counts in FragPipe only include those from MS/MS. If only counting the proteins from MS/MS, MaxQuant has 1012 razor+unique, and FragPipe has 950 razor+unique. Then, I have the following questions.

I see that you have 12 runs starting with "S", three runs starting with "G", three runs starting with "H", and 20 runs starting with "lib". For the "S", "G", and "H", MS/MS gets less IDs than MBR. For the "lib", MS/MS gets more IDs than MBR. I guess you are analyzing single-cell-like data which relies on "lib" runs to contribute IDs to single cell runs. If my guess is correct, could you please help to check

  1. What is the MS/MS analyzer of your data? Are all of those runs from linear ion trap or orbitrap?
  2. Do all runs have the same gradient length?
  3. To make a fair comparison, you need to set MBR ion, peptide, and protein fdr to 1 (which equals 100%) because MaxQuant does not have any FDR control for MBR. Those additional proteins from MaxQuant's MBR might be false positives.

Best,

Fengchao

On Wed, Sep 1, 2021 at 9:04 PM huangyingzhang @.***> wrote:

Hi, Dr. Yu,

Yes. There are more protein groups from FragPipe. But what I am focus on is the protein group identified in individual sample other than total proteins. For example, There is 2348 proteins with unique peptides in sample named G-1 from Maxquant proteinGroups result, while the same sample in FragPipe is 478 proteins with Unique Spectral count and 200 proteins with Unique intensity.

Regards, Yingzhang

On 1 Sep 2021, at 11:16 PM, Fengchao @.**@.>> wrote:

Thanks for your file.

In your proteinGroups.txt from MaxQuant, there are 11605 protein groups after filtering out "Only identified by site", "Reverse", and "Potential contaminant".

In your combined_protein.tsv from FragPipe, there are 13020 proteins.

To me, FragPipe found more protein than MaxQuant. Can you double check and elaborate more about your issue?

Best,

Fengchao

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub< https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-910385104>, or unsubscribe< https://github.com/notifications/unsubscribe-auth/AKLH75IQ3VVL63ORBEGDZE3T7Y7TZANCNFSM5DD5GK5A>.

Triage notifications on the go with GitHub Mobile for iOS< https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android< https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-910979832, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABU27W747LKC5ENZZC6DNNLT73ESPANCNFSM5DD5GK5A . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

-- Dr. Fengchao Yu Research Investigator University of Michigan

huangyingzhang commented 3 years ago

Hi, Dr. Yu,

Yes, exactly what you said. I am doing nanoproteomics. Samples starting with “S”,”G”,”H” rely on “lib”.

  1. All of those runs with orbitrap (Orbitrap FusionTM LumosTM TribridTM (Thermo Fisher Scientific, San Jose, CA)
  2. All of them with the same gradient length.
  3. For third one, I may need to run Maxquant again.

What I am confused is which one is the MBR result. Is the “Total Spectral count” ( 1162 proteins for sample G-1) ? Thanks.

Regards, Yingzhang

On 2 Sep 2021, at 9:42 AM, Fengchao @.**@.>> wrote:

Hi Yingzhang,

First of all, the unique peptides in MaxQuant include those from MS/MS and MBR, while the unique spectral counts in FragPipe only include those from MS/MS. If only counting the proteins from MS/MS, MaxQuant has 1012 razor+unique, and FragPipe has 950 razor+unique. Then, I have the following questions.

I see that you have 12 runs starting with "S", three runs starting with "G", three runs starting with "H", and 20 runs starting with "lib". For the "S", "G", and "H", MS/MS gets less IDs than MBR. For the "lib", MS/MS gets more IDs than MBR. I guess you are analyzing single-cell-like data which relies on "lib" runs to contribute IDs to single cell runs. If my guess is correct, could you please help to check

  1. What is the MS/MS analyzer of your data? Are all of those runs from linear ion trap or orbitrap?
  2. Do all runs have the same gradient length?
  3. To make a fair comparison, you need to set MBR ion, peptide, and protein fdr to 1 (which equals 100%) because MaxQuant does not have any FDR control for MBR. Those additional proteins from MaxQuant's MBR might be false positives.

Best,

Fengchao

On Wed, Sep 1, 2021 at 9:04 PM huangyingzhang @.***> wrote:

Hi, Dr. Yu,

Yes. There are more protein groups from FragPipe. But what I am focus on is the protein group identified in individual sample other than total proteins. For example, There is 2348 proteins with unique peptides in sample named G-1 from Maxquant proteinGroups result, while the same sample in FragPipe is 478 proteins with Unique Spectral count and 200 proteins with Unique intensity.

Regards, Yingzhang

On 1 Sep 2021, at 11:16 PM, Fengchao @.**@.>> wrote:

Thanks for your file.

In your proteinGroups.txt from MaxQuant, there are 11605 protein groups after filtering out "Only identified by site", "Reverse", and "Potential contaminant".

In your combined_protein.tsv from FragPipe, there are 13020 proteins.

To me, FragPipe found more protein than MaxQuant. Can you double check and elaborate more about your issue?

Best,

Fengchao

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub< https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-910385104>, or unsubscribe< https://github.com/notifications/unsubscribe-auth/AKLH75IQ3VVL63ORBEGDZE3T7Y7TZANCNFSM5DD5GK5A>.

Triage notifications on the go with GitHub Mobile for iOS< https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android< https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-910979832, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABU27W747LKC5ENZZC6DNNLT73ESPANCNFSM5DD5GK5A . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

-- Dr. Fengchao Yu Research Investigator University of Michigan

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/Nesvilab/FragPipe/issues/449#issuecomment-911014425, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AKLH75OTAUFH4VWJJAJJSV3T73JAXANCNFSM5DD5GK5A. Triage notifications on the go with GitHub Mobile for iOShttps://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Androidhttps://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

fcyu commented 3 years ago

Hi Yingzhang,

Thanks for your information. For the third one, you need to rerun FragPipe, not MaxQuant.

Everything related to spectral count is from MS/MS.

I will also analyze your data with FragPipe to see what is going on. In the meantime, could you upload all files in MaxQuant's combine/txt folder and all fasta files?

Thanks,

Fengchao

On Wed, Sep 1, 2021 at 10:28 PM huangyingzhang @.***> wrote:

Hi, Dr. Yu,

Yes, exactly what you said. I am doing nanoproteomics. Samples starting with “S”,”G”,”H” rely on “lib”.

  1. All of those runs with orbitrap (Orbitrap FusionTM LumosTM TribridTM (Thermo Fisher Scientific, San Jose, CA)
  2. All of them with the same gradient length.
  3. For third one, I may need to run Maxquant again.

What I am confused is which one is the MBR result. Is the “Total Spectral count” ( 1162 proteins for sample G-1) ? Thanks.

Regards, Yingzhang

On 2 Sep 2021, at 9:42 AM, Fengchao @.**@.>> wrote:

Hi Yingzhang,

First of all, the unique peptides in MaxQuant include those from MS/MS and MBR, while the unique spectral counts in FragPipe only include those from MS/MS. If only counting the proteins from MS/MS, MaxQuant has 1012 razor+unique, and FragPipe has 950 razor+unique. Then, I have the following questions.

I see that you have 12 runs starting with "S", three runs starting with "G", three runs starting with "H", and 20 runs starting with "lib". For the "S", "G", and "H", MS/MS gets less IDs than MBR. For the "lib", MS/MS gets more IDs than MBR. I guess you are analyzing single-cell-like data which relies on "lib" runs to contribute IDs to single cell runs. If my guess is correct, could you please help to check

  1. What is the MS/MS analyzer of your data? Are all of those runs from linear ion trap or orbitrap?
  2. Do all runs have the same gradient length?
  3. To make a fair comparison, you need to set MBR ion, peptide, and protein fdr to 1 (which equals 100%) because MaxQuant does not have any FDR control for MBR. Those additional proteins from MaxQuant's MBR might be false positives.

Best,

Fengchao

On Wed, Sep 1, 2021 at 9:04 PM huangyingzhang @.***> wrote:

Hi, Dr. Yu,

Yes. There are more protein groups from FragPipe. But what I am focus on is the protein group identified in individual sample other than total proteins. For example, There is 2348 proteins with unique peptides in sample named G-1 from Maxquant proteinGroups result, while the same sample in FragPipe is 478 proteins with Unique Spectral count and 200 proteins with Unique intensity.

Regards, Yingzhang

On 1 Sep 2021, at 11:16 PM, Fengchao @.**@.>> wrote:

Thanks for your file.

In your proteinGroups.txt from MaxQuant, there are 11605 protein groups after filtering out "Only identified by site", "Reverse", and "Potential contaminant".

In your combined_protein.tsv from FragPipe, there are 13020 proteins.

To me, FragPipe found more protein than MaxQuant. Can you double check and elaborate more about your issue?

Best,

Fengchao

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub< https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-910385104>,

or unsubscribe<

https://github.com/notifications/unsubscribe-auth/AKLH75IQ3VVL63ORBEGDZE3T7Y7TZANCNFSM5DD5GK5A>.

Triage notifications on the go with GitHub Mobile for iOS<

https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>

or Android<

https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-910979832,

or unsubscribe < https://github.com/notifications/unsubscribe-auth/ABU27W747LKC5ENZZC6DNNLT73ESPANCNFSM5DD5GK5A>

. Triage notifications on the go with GitHub Mobile for iOS < https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>

or Android < https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

-- Dr. Fengchao Yu Research Investigator University of Michigan

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub< https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-911014425>, or unsubscribe< https://github.com/notifications/unsubscribe-auth/AKLH75OTAUFH4VWJJAJJSV3T73JAXANCNFSM5DD5GK5A>.

Triage notifications on the go with GitHub Mobile for iOS< https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android< https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-911054998, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABU27W2ZKLSFPNPDYFSG45DT73ONBANCNFSM5DD5GK5A . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

-- Dr. Fengchao Yu Research Investigator University of Michigan

huangyingzhang commented 3 years ago

Hi, Dr. Yu,

Thank you so much. I have uploaded the fasta file and will upload Maxuquants combine /txt later.

Regards, Yingzhang

On 2 Sep 2021, at 10:32 AM, Fengchao @.**@.>> wrote:

Hi Yingzhang,

Thanks for your information. For the third one, you need to rerun FragPipe, not MaxQuant.

Everything related to spectral count is from MS/MS.

I will also analyze your data with FragPipe to see what is going on. In the meantime, could you upload all files in MaxQuant's combine/txt folder and all fasta files?

Thanks,

Fengchao

On Wed, Sep 1, 2021 at 10:28 PM huangyingzhang @.***> wrote:

Hi, Dr. Yu,

Yes, exactly what you said. I am doing nanoproteomics. Samples starting with “S”,”G”,”H” rely on “lib”.

  1. All of those runs with orbitrap (Orbitrap FusionTM LumosTM TribridTM (Thermo Fisher Scientific, San Jose, CA)
  2. All of them with the same gradient length.
  3. For third one, I may need to run Maxquant again.

What I am confused is which one is the MBR result. Is the “Total Spectral count” ( 1162 proteins for sample G-1) ? Thanks.

Regards, Yingzhang

On 2 Sep 2021, at 9:42 AM, Fengchao @.**@.>> wrote:

Hi Yingzhang,

First of all, the unique peptides in MaxQuant include those from MS/MS and MBR, while the unique spectral counts in FragPipe only include those from MS/MS. If only counting the proteins from MS/MS, MaxQuant has 1012 razor+unique, and FragPipe has 950 razor+unique. Then, I have the following questions.

I see that you have 12 runs starting with "S", three runs starting with "G", three runs starting with "H", and 20 runs starting with "lib". For the "S", "G", and "H", MS/MS gets less IDs than MBR. For the "lib", MS/MS gets more IDs than MBR. I guess you are analyzing single-cell-like data which relies on "lib" runs to contribute IDs to single cell runs. If my guess is correct, could you please help to check

  1. What is the MS/MS analyzer of your data? Are all of those runs from linear ion trap or orbitrap?
  2. Do all runs have the same gradient length?
  3. To make a fair comparison, you need to set MBR ion, peptide, and protein fdr to 1 (which equals 100%) because MaxQuant does not have any FDR control for MBR. Those additional proteins from MaxQuant's MBR might be false positives.

Best,

Fengchao

On Wed, Sep 1, 2021 at 9:04 PM huangyingzhang @.***> wrote:

Hi, Dr. Yu,

Yes. There are more protein groups from FragPipe. But what I am focus on is the protein group identified in individual sample other than total proteins. For example, There is 2348 proteins with unique peptides in sample named G-1 from Maxquant proteinGroups result, while the same sample in FragPipe is 478 proteins with Unique Spectral count and 200 proteins with Unique intensity.

Regards, Yingzhang

On 1 Sep 2021, at 11:16 PM, Fengchao @.**@.>> wrote:

Thanks for your file.

In your proteinGroups.txt from MaxQuant, there are 11605 protein groups after filtering out "Only identified by site", "Reverse", and "Potential contaminant".

In your combined_protein.tsv from FragPipe, there are 13020 proteins.

To me, FragPipe found more protein than MaxQuant. Can you double check and elaborate more about your issue?

Best,

Fengchao

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub< https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-910385104>,

or unsubscribe<

https://github.com/notifications/unsubscribe-auth/AKLH75IQ3VVL63ORBEGDZE3T7Y7TZANCNFSM5DD5GK5A>.

Triage notifications on the go with GitHub Mobile for iOS<

https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>

or Android<

https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-910979832,

or unsubscribe < https://github.com/notifications/unsubscribe-auth/ABU27W747LKC5ENZZC6DNNLT73ESPANCNFSM5DD5GK5A>

. Triage notifications on the go with GitHub Mobile for iOS < https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>

or Android < https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

-- Dr. Fengchao Yu Research Investigator University of Michigan

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub< https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-911014425>, or unsubscribe< https://github.com/notifications/unsubscribe-auth/AKLH75OTAUFH4VWJJAJJSV3T73JAXANCNFSM5DD5GK5A>.

Triage notifications on the go with GitHub Mobile for iOS< https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android< https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-911054998, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABU27W2ZKLSFPNPDYFSG45DT73ONBANCNFSM5DD5GK5A . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

-- Dr. Fengchao Yu Research Investigator University of Michigan

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/Nesvilab/FragPipe/issues/449#issuecomment-911057733, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AKLH75PCYYQF4NKLTCLIBYTT73O2VANCNFSM5DD5GK5A. Triage notifications on the go with GitHub Mobile for iOShttps://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Androidhttps://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

huangyingzhang commented 3 years ago

Dear Dr. Yu,

I rerun LFQ-MBR with MBR ion FDR =1, MBR peptide FDR=1, MBR protein FDR=1( the default FDR of peptide and protein is 1).

For sample G_1, in combined_protein.tsv, the protein number with the ’total intensity’>0 is 2258, while inside the G_1 directory, in protein.tsv, the number is 4562. With MBR ion FDR =0.01, in combined_protein.tsv, the number is 636; in protein.tsv, the number is 1138. I also run with MBR ion FDR =0.1, in combined_protein.tsv, the number is 1807, while in protein.tsv, the number is 3377.

It seems that the MBR ion FDR indeed have big effect on the result. I have anther question. Why the result between combined_protein.tsv and individual sample protein.tsv is different? Thanks.

Regards, Yingzhang

On 2 Sep 2021, at 10:32 AM, Fengchao @.**@.>> wrote:

Hi Yingzhang,

Thanks for your information. For the third one, you need to rerun FragPipe, not MaxQuant.

Everything related to spectral count is from MS/MS.

I will also analyze your data with FragPipe to see what is going on. In the meantime, could you upload all files in MaxQuant's combine/txt folder and all fasta files?

Thanks,

Fengchao

On Wed, Sep 1, 2021 at 10:28 PM huangyingzhang @.***> wrote:

Hi, Dr. Yu,

Yes, exactly what you said. I am doing nanoproteomics. Samples starting with “S”,”G”,”H” rely on “lib”.

  1. All of those runs with orbitrap (Orbitrap FusionTM LumosTM TribridTM (Thermo Fisher Scientific, San Jose, CA)
  2. All of them with the same gradient length.
  3. For third one, I may need to run Maxquant again.

What I am confused is which one is the MBR result. Is the “Total Spectral count” ( 1162 proteins for sample G-1) ? Thanks.

Regards, Yingzhang

On 2 Sep 2021, at 9:42 AM, Fengchao @.**@.>> wrote:

Hi Yingzhang,

First of all, the unique peptides in MaxQuant include those from MS/MS and MBR, while the unique spectral counts in FragPipe only include those from MS/MS. If only counting the proteins from MS/MS, MaxQuant has 1012 razor+unique, and FragPipe has 950 razor+unique. Then, I have the following questions.

I see that you have 12 runs starting with "S", three runs starting with "G", three runs starting with "H", and 20 runs starting with "lib". For the "S", "G", and "H", MS/MS gets less IDs than MBR. For the "lib", MS/MS gets more IDs than MBR. I guess you are analyzing single-cell-like data which relies on "lib" runs to contribute IDs to single cell runs. If my guess is correct, could you please help to check

  1. What is the MS/MS analyzer of your data? Are all of those runs from linear ion trap or orbitrap?
  2. Do all runs have the same gradient length?
  3. To make a fair comparison, you need to set MBR ion, peptide, and protein fdr to 1 (which equals 100%) because MaxQuant does not have any FDR control for MBR. Those additional proteins from MaxQuant's MBR might be false positives.

Best,

Fengchao

On Wed, Sep 1, 2021 at 9:04 PM huangyingzhang @.***> wrote:

Hi, Dr. Yu,

Yes. There are more protein groups from FragPipe. But what I am focus on is the protein group identified in individual sample other than total proteins. For example, There is 2348 proteins with unique peptides in sample named G-1 from Maxquant proteinGroups result, while the same sample in FragPipe is 478 proteins with Unique Spectral count and 200 proteins with Unique intensity.

Regards, Yingzhang

On 1 Sep 2021, at 11:16 PM, Fengchao @.**@.>> wrote:

Thanks for your file.

In your proteinGroups.txt from MaxQuant, there are 11605 protein groups after filtering out "Only identified by site", "Reverse", and "Potential contaminant".

In your combined_protein.tsv from FragPipe, there are 13020 proteins.

To me, FragPipe found more protein than MaxQuant. Can you double check and elaborate more about your issue?

Best,

Fengchao

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub< https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-910385104>,

or unsubscribe<

https://github.com/notifications/unsubscribe-auth/AKLH75IQ3VVL63ORBEGDZE3T7Y7TZANCNFSM5DD5GK5A>.

Triage notifications on the go with GitHub Mobile for iOS<

https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>

or Android<

https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-910979832,

or unsubscribe < https://github.com/notifications/unsubscribe-auth/ABU27W747LKC5ENZZC6DNNLT73ESPANCNFSM5DD5GK5A>

. Triage notifications on the go with GitHub Mobile for iOS < https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>

or Android < https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

-- Dr. Fengchao Yu Research Investigator University of Michigan

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub< https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-911014425>, or unsubscribe< https://github.com/notifications/unsubscribe-auth/AKLH75OTAUFH4VWJJAJJSV3T73JAXANCNFSM5DD5GK5A>.

Triage notifications on the go with GitHub Mobile for iOS< https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android< https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/449#issuecomment-911054998, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABU27W2ZKLSFPNPDYFSG45DT73ONBANCNFSM5DD5GK5A . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

-- Dr. Fengchao Yu Research Investigator University of Michigan

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/Nesvilab/FragPipe/issues/449#issuecomment-911057733, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AKLH75PCYYQF4NKLTCLIBYTT73O2VANCNFSM5DD5GK5A. Triage notifications on the go with GitHub Mobile for iOShttps://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Androidhttps://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

fcyu commented 3 years ago

Hi Yingzhang,

I also analyze your data and reproduce what you found. It looks like there are a lot of false positives if without MBR ion FDR < 0.01 as a filter. The target and decoy MBR ion histograms from G_1 also verify it (see below figure). Since MaxQuant's MBR does not have any FDR control, this is why it gets more proteins than IonQuant with MBR FDR < 0.01.

image

Also, the peptides in your sample seems not having many replicates. Using min ion = 1 gives much more proteins with non-zero intensities than using min ion = 2.

Regarding the difference between combined_protein.tsv and protein.tsv, the intensities in combined_protein are from MaxLFQ algorithm and min ion = 2, while the intensities in protein.tsv are from top N and no min ion filtering. Since your peptides do not have may replicates, there are more non-zero intensity proteins in protein.tsv.

Best,

Fengchao

huangyingzhang commented 3 years ago

Hi, Dr. Yu,

Many many thanks. Will try min ion=1.

Best, Yingzhang

On 4 Sep 2021, at 9:40 PM, Fengchao @.**@.>> wrote:

Hi Yingzhang,

I also analyze your data and reproduce what you found. It looks like there are a lot of false positives if without MBR ion FDR < 0.01 as a filter. The target and decoy MBR ion histograms from G_1 also verify it (see below figure). Since MaxQuant's MBR does not have any FDR control, this is why it gets more proteins than IonQuant with MBR FDR < 0.01.

[image]https://user-images.githubusercontent.com/6926299/132096251-703fc739-485a-4233-a238-52e8af6b6152.png

Also, the peptides in your sample seems not having many replicates. Using min ion = 1 gives much more proteins with non-zero intensities than using min ion = 2.

Regarding the difference between combined_protein.tsv and protein.tsv, the intensities in combined_protein are from MaxLFQ algorithm and min ion = 2, while the intensities in protein.tsv are from top N and no min ion filtering. Since your peptides do not have may replicates, there are more non-zero intensity proteins in protein.tsv.

Best,

Fengchao

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/Nesvilab/FragPipe/issues/449#issuecomment-912976132, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AKLH75O6M5UHZXRRIQ2RQOTUAIOTDANCNFSM5DD5GK5A. Triage notifications on the go with GitHub Mobile for iOShttps://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Androidhttps://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.