Libra quantification on TMT 16-plex returns zeros only

31 views
Skip to first unread message

Yasir Ahmed

unread,
Sep 8, 2022, 3:02:54 PM9/8/22
to spctools-discuss
Greetings,

I'm having difficulty performing TMT quantification using Libra on a 16-plex dataset. Basically, the "quantification.tsv" file that Libra outputs only contains zeros for both peptides and proteins. The Comet search runs just fine. I'm at a loss for where things are going wrong. The Mass Spec facility used ProteomeDiscoverer for the quantification, and it worked fine so the issue is not the raw data.

My TPP run repository can be found here, and the "condition.xml" file for Libra—along with the Comet params file—is in the "params" folder. The commands used to run the TPP processes can be found in the file called "Snakefile". I'm happy to provide any additional information needed to help troubleshoot this. Any pointers/advise would be greatly appreciated.

Cheers,
Yasir

Luis Mendoza

unread,
Sep 8, 2022, 3:45:22 PM9/8/22
to spctools...@googlegroups.com
Hello Yasir,

Can you try re-running Libra without the centroiding option?  (i.e. set type="0") Or are your mzML data not centroided? 
Another thing to try: are you able to open the PeptideProphet + Libra results in the PepXMLViewer?  You should be able to see if there are a significant number of quantified PSMs -- you should also be able to examine a few individual spectra and zoom in on the reporter ion signals.
Lastly, what version of TPP are you running?

Cheers,
--Luis


--
You received this message because you are subscribed to the Google Groups "spctools-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to spctools-discu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/spctools-discuss/7cb42706-53ae-4c94-acd2-ea0293fea8b2n%40googlegroups.com.

Yasir Ahmed

unread,
Sep 8, 2022, 5:08:36 PM9/8/22
to spctools...@googlegroups.com
Hi Luis,

Thanks a lot for the response. 

Can you try re-running Libra without the centroiding option?  (i.e. set type="0") Or are your mzML data not centroided? 

I just tried that, but no difference in outcome.

Another thing to try: are you able to open the PeptideProphet + Libra results in the PepXMLViewer?  You should be able to see if there are a significant number of quantified PSMs -- you should also be able to examine a few individual spectra and zoom in on the reporter ion signals.

I’m able to look at a few individual spectra, and I think they all look mostly good—but frankly I’m not too sure how to evaluate their suitability (see example screenshots below). 

Lastly, what version of TPP are you running?

6.1.0

Thanks again!
Yasir





Cheers,
--Luis


On Thu, Sep 8, 2022 at 12:02 PM Yasir Ahmed <yaza...@gmail.com> wrote:
Greetings,

I'm having difficulty performing TMT quantification using Libra on a 16-plex dataset. Basically, the "quantification.tsv" file that Libra outputs only contains zeros for both peptides and proteins. The Comet search runs just fine. I'm at a loss for where things are going wrong. The Mass Spec facility used ProteomeDiscoverer for the quantification, and it worked fine so the issue is not the raw data.

My TPP run repository can be found here, and the "condition.xml" file for Libra—along with the Comet params file—is in the "params" folder. The commands used to run the TPP processes can be found in the file called "Snakefile". I'm happy to provide any additional information needed to help troubleshoot this. Any pointers/advise would be greatly appreciated.

Cheers,
Yasir

--
You received this message because you are subscribed to the Google Groups "spctools-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to spctools-discu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/spctools-discuss/7cb42706-53ae-4c94-acd2-ea0293fea8b2n%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "spctools-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to spctools-discu...@googlegroups.com.

Yasir Ahmed

unread,
Sep 8, 2022, 5:19:24 PM9/8/22
to spctools...@googlegroups.com
Hi Luis,

Just to follow up: I found a handful of cases where one of the channels has super high quantification and observable reporter ions (see screenshots below), albeit it should be way more than this, I believe. I’ll keep digging, but let me know if you have any other ideas.

Again, thanks for you suggestions.

Cheers,
Yasir





On Sep 8, 2022, at 5:07 PM, Yasir Ahmed <yaza...@gmail.com> wrote:

Hi Luis,

Thanks a lot for the response. 

Can you try re-running Libra without the centroiding option?  (i.e. set type="0") Or are your mzML data not centroided? 

I just tried that, but no difference in outcome.

Another thing to try: are you able to open the PeptideProphet + Libra results in the PepXMLViewer?  You should be able to see if there are a significant number of quantified PSMs -- you should also be able to examine a few individual spectra and zoom in on the reporter ion signals.

I’m able to look at a few individual spectra, and I think they all look mostly good—but frankly I’m not too sure how to evaluate their suitability (see example screenshots below). 

Lastly, what version of TPP are you running?

6.1.0

Thanks again!
Yasir

<Screen Shot 2022-09-08 at 5.00.50 PM.png>

<Screen Shot 2022-09-08 at 5.01.42 PM.png>

Eric Deutsch

unread,
Sep 8, 2022, 5:20:58 PM9/8/22
to spctools...@googlegroups.com, Eric Deutsch

I don’t see any signal down where the reporter ions are at ~120 m/z

 

In fact, that looks a lot like an ion trap spectrum, which typically don’t have signal down at ~120.

Eric Deutsch

unread,
Sep 8, 2022, 5:28:11 PM9/8/22
to spctools...@googlegroups.com, Eric Deutsch

I’m thinking that ion trap spectra are usually unsuitable for TMT:

 

 

I think you would want to acquire fragmentation spectra with Orbitrap HCD?

 

 

From: spctools...@googlegroups.com <spctools...@googlegroups.com> On Behalf Of Yasir Ahmed
Sent: Thursday, September 8, 2022 2:19 PM
To: spctools...@googlegroups.com
Subject: Re: [spctools-discuss] Libra quantification on TMT 16-plex returns zeros only

 

Hi Luis,

Luis Mendoza

unread,
Sep 8, 2022, 6:16:07 PM9/8/22
to spctools...@googlegroups.com
Hi Yasir,
Is your data Thermo SPS where the reporter ions are in MS3?  If so, please check the relevant box when producing the condition file:

image.png


As Eric noted, it does not seem like your MS2 spectra contain reporter ion signals, at least from the one spectrum you posted.  Are there others where this is not the case?

Cheers,
--Luis


Yasir Ahmed

unread,
Sep 9, 2022, 4:54:30 PM9/9/22
to spctools...@googlegroups.com
Thanks Eric and Luis! Turns out I was using the wrong/older syntax in the condition file (targetMS level = “3”). The quantification works now. Thanks for your help!!!

Cheers,
Yasir

On Sep 8, 2022, at 6:15 PM, 'Luis Mendoza' via spctools-discuss <spctools...@googlegroups.com> wrote:

Hi Yasir,
Is your data Thermo SPS where the reporter ions are in MS3?  If so, please check the relevant box when producing the condition file:

<image.png>


As Eric noted, it does not seem like your MS2 spectra contain reporter ion signals, at least from the one spectrum you posted.  Are there others where this is not the case?

Cheers,
--Luis

On Thu, Sep 8, 2022 at 2:28 PM 'Eric Deutsch' via spctools-discuss <spctools...@googlegroups.com> wrote:

I’m thinking that ion trap spectra are usually unsuitable for TMT:

 

<image001.png>

 

I think you would want to acquire fragmentation spectra with Orbitrap HCD?

 
 

From: spctools...@googlegroups.com <spctools...@googlegroups.com> On Behalf Of Yasir Ahmed
Sent: Thursday, September 8, 2022 2:19 PM
To: spctools...@googlegroups.com
Subject: Re: [spctools-discuss] Libra quantification on TMT 16-plex returns zeros only

 

Hi Luis,

 

Just to follow up: I found a handful of cases where one of the channels has super high quantification and observable reporter ions (see screenshots below), albeit it should be way more than this, I believe. I’ll keep digging, but let me know if you have any other ideas.

 

Again, thanks for you suggestions.

 

Cheers,

Yasir

 

<image002.png>

 

<image003.png>

 

<image004.png>

Luis Mendoza

unread,
Sep 9, 2022, 5:54:23 PM9/9/22
to spctools...@googlegroups.com
Great, glad that it is now working for you!  We will remove references to that deprecated parameter from our documentation.
Cheers,
--Luis


Reply all
Reply to author
Forward
0 new messages