TMT Integrator: Ratio to Abundance Conversion options

[enonimos]

Hello, I have been using the isobaric quant tool in FragPipe 17 and now version 18, and I wanted to report an unexpected result when using different "ratio to abundance" conversion options. I have routinely used the tool with the default "Use MS1 Intensity" and "Top 3" options enabled, plus I also like the option to display the Reference Intensity channels, so I enable that option. For these options, I noticed that the number of valid values in the Ref/pool channels is 100-200 less than valid values in the corresponding reporter channels. I thought there might be an experimental issue with the datasets, but if I disable "Use MS1 intensity" then the valid values in the RefInt columns is the same or nearly the same as their corresponding reporter columns. The latter result is what I would expect from an experimental perspective. Notably, the # of valid values in the primary report channels is exactly the same regardless of the "Use MS1 Intensity" status. This is what I had expected since the "Use MS1 Intensity" is just used for ratio conversion and shouldn't change the underlying quant.

Taking this together, I am wondering why the RefInt column seem to be affected?

And as a general observation about ratio conversion, I noticed that when "Use MS1 intensity" is disabled (and presumably the MS2 intensities are used), TMT Integrator still performs FreeQuant MS1 LFQ. I'm wondering if this is still needed in this case?

Thank you in advance for your help in clarifying this observation. And overall, thank you for the recent updates, really enjoying v18.

Todd

[huiyinc]

Hi Todd,

Thanks very much for sharing. Is it possible for you to share the PSM tables with us (one generated using FragPipe v17 and another generated using FragPipe v18) along with the fasta file? I would like to check if the values in the two PSM tables are different. If they are the same, then I will have to check TMT-Integrator to figure out the reason of having different valid values.

Also, thanks for the suggestion. Yes, TMT-Integrator doesn't need the information of MS1 intensity when "Use MS1 intensity" is disabled. FreeQuant MS1 LFQ is run by Philosopher, and I personally think it's good to have MS1 intensity in PSM tables even though TMT-Integrator doesn't need it.

Please feel free to ask if there is any question.

Best,

Huiyin

enonimos @.***> 於 2022年6月11日 上午10:49 寫道：



Hello, I have been using the isobaric quant tool in FragPipe 17 and now version 18, and I wanted to report an unexpected result when using different "ratio to abundance" conversion options. I have routinely used the tool with the default "Use MS1 Intensity" and "Top 3" options enabled, plus I also like the option to display the Reference Intensity channels, so I enable that option. For these options, I noticed that the number of valid values in the Ref/pool channels is 100-200 less than valid values in the corresponding reporter channels. I thought there might be an experimental issue with the datasets, but if I disable "Use MS1 intensity" then the valid values in the RefInt columns is the same or nearly the same as their corresponding reporter columns. The latter result is what I would expect from an experimental perspective. Notably, the # of valid values in the primary report channels is exactly the same regardless of the "Use MS1 Intensity" status. This is what I had expected since the "Use MS1 Intensity" is just used for ratio conversion and shouldn't change the underlying quant.

Taking this together, I am wondering why the RefInt column seem to be affected?

And as a general observation about ratio conversion, I noticed that when "Use MS1 intensity" is disabled (and presumably the MS2 intensities are used), TMT Integrator still performs FreeQuant MS1 LFQ. I'm wondering if this is still needed in this case?

Thank you in advance for your help in clarifying this observation. And overall, thank you for the recent updates, really enjoying v18.

Todd

— Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/714, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALAWWA24QK6COGJ3GKJXTC3VOP5FJANCNFSM5YPK635Q . You are receiving this because you are subscribed to this thread.Message ID: @.***>

[danielgeiszler]

@huiyinc Will TMT-I expect the purity column when FreeQuant is disabled?

[huiyinc]

@danielgeiszler Yes, the purity column is needed. I thought the purity is calculated when labelquant is performed.

Huiyin

danielgeiszler @.***> 於 2022年6月13日 週一 下午12:18寫道：

@huiyinc https://github.com/huiyinc Will TMT-I expect the purity column when FreeQuant is disabled?

— Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/714#issuecomment-1153448707, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALAWWAZPEOIYQARVRMTSODLVO2ZADANCNFSM5YPK635Q . You are receiving this because you were mentioned.Message ID: @.***>

-- Hui-Yin Chang, 張彙音 Assistant Professor Department of Biomedical Sciences and Engineering National Central University, Taiwan

[danielgeiszler]

@huiyinc I think it's connected to FreeQuant, not LabelQuant. I was running some tests with it a few days ago and wasn't using LabelQuant.

@prvst would know.

[huiyinc]

@danielgeiszler If so, FreeQuant is needed for TMT-I to have the purity column in PSM tables. Are you running FreeQuant alone using the command line?

Huiyin

danielgeiszler @.***> 於 2022年6月13日 週一 下午1:05寫道：

@huiyinc https://github.com/huiyinc I think it's connected to FreeQuant, not LabelQuant. I was running some tests with it a few days ago and wasn't using LabelQuant.

— Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/714#issuecomment-1153476284, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALAWWAZNBGT3M32XSRZZ7G3VO26TLANCNFSM5YPK635Q . You are receiving this because you were mentioned.Message ID: @.***>

-- Hui-Yin Chang, 張彙音 Assistant Professor Department of Biomedical Sciences and Engineering National Central University, Taiwan

[prvst]

both the label-free, and isobaric quantification will add the purity information to the PSM table.

[enonimos]

@huiyinc The PSM tables between v17 and v18 are identical, and the behavior I observed is the same between v17 and v18, so this does appear to be a TMT-Integrator and not an issue created by v18 specifically.

Thanks to everyone for commentary about benefit/need to still have FreeQuant even with MS2-based ratio conversion.

[huiyinc]

Hi Todd,

Can you please send me the psm tables and fasta files? I want to check what makes the difference. Thanks.

Huiyin

enonimos @.***> 於 2022年6月14日 週二 上午4:19寫道：

@huiyinc https://github.com/huiyinc The PSM tables between v17 and v18 are identical, and the behavior I observed is the same between v17 and v18, so this does appear to be a TMT-Integrator and not an issue created by v18 specifically.

Thanks to everyone for commentary about benefit/need to still have FreeQuant even with MS2-based ratio conversion.

— Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/714#issuecomment-1154388381, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALAWWA7DHGM55QVQHYDCPA3VO6JTPANCNFSM5YPK635Q . You are receiving this because you were mentioned.Message ID: @.***>

-- Hui-Yin Chang, 張彙音 Assistant Professor Department of Biomedical Sciences and Engineering National Central University, Taiwan

[enonimos]

Hi Huiyin, Here is a link to a zip archive that has the files you requested. https://www.dropbox.com/s/39s9mojtr2fdio3/TMT%20Integrator%20MS1%20vs%20MS2%20files.zip?dl=0

The two PSM files are the same dataset analyzed by TMT Integrator with MS1 Intensity box enabled (MS1) or disabled (MS2). With my casual inspection, I didn't see any differences, but hopefully they are helpful. Since I still see a difference in the corresponding protein level tmt-report output, I have included these for your reference.

Thank you for troubleshooting and let me know if you were able to download the files successfully.

Todd

[huiyinc]

Hi Todd,

I notice that the two PSM tables are generated using FragPipe v18, psm_WT_2M_01_v18_MS1_Int.tsv and psm_WT_2M_01_v18_MS2_Int.tsv. Can you please tell me how the two tables are generated using FragPipe? What's the differences in the FragPipe workflow?

For your question, TMT-Integrator filters out PSMs with the pool intensity of 0, but doesn't filter out those with MS1 intensity of 0. That's why you have missing values (i.e., NA) in the reference channels of TMT-Integrator reports when "Use MS1 Intensity" is enabled. You may find out that the ratios between using and not using MS1 intensity are the same.

Hope this answers your question. Please feel free to ask if you have any questions.

Best,

Huiyin

enonimos @.***> 於 2022年6月15日 週三 上午9:16寫道：

Hi Huiyin, Here is a link to a zip archive that has the files you requested. https://www.dropbox.com/s/39s9mojtr2fdio3/TMT%20Integrator%20MS1%20vs%20MS2%20files.zip?dl=0

The two PSM files are the same dataset analyzed by TMT Integrator with MS1 Intensity box enabled (MS1) or disabled (MS2). With my casual inspection, I didn't see any differences, but hopefully they are helpful. Since I still see a difference in the corresponding protein level tmt-report output, I have included these for your reference.

Thank you for troubleshooting and let me know if you were able to download the files successfully.

Todd

— Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/714#issuecomment-1155870119, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALAWWAZAATNZNA2MZ3EYK3TVPEVHXANCNFSM5YPK635Q . You are receiving this because you were mentioned.Message ID: @.***>

-- Hui-Yin Chang, 張彙音 Assistant Professor Department of Biomedical Sciences and Engineering National Central University, Taiwan

[enonimos]

Hi Huiyin, Yes that's correct, I sent the PSM tables for v18 because the issue I saw isn't one that appeared after the update, it occurs in both v17 and v18 in the same way. The difference in FragPipe workflow is only in isobaric tab, "Use MS1 Intensity" was either enabled (MS1_Int file) or disabled (MS2_Int file).

In the TMT reports, the experimental channels have the same number of quantitative values independent of the MS1 intensity option, and as you said, the ratios are the same, while the abundances only differ systematically I assume due to different scaling fator. This all makes sense to me.

So the issue I raised only affects the RefInt/Pool channel column. If I understand what your saying with TMT-I filtering, there is the original reference channel coming from MS2, then if Use MS1 Intensity is enabled, the MS2 reference values are overwritten by the new MS1 values. However, if the MS1 intensity value is 0, then an NA will be inserted instead, and that's why I have more missing values for those columns.

If this is correct, then I have a follow-up question. How are the ratios converted to abundances for those case where MS1 intensity calculation fails? Does it default back to the MS2 intensity of the reference channel? If so, then shouldn't those values be listed in the report instead of NA?

Thank you in advance for the clarification.

Todd

[huiyinc]

Hi Todd,

Thanks for the explanation. Yes, enabling or disabling "Use MS1 Intensity" in TMT-Integrator doesn't affect the generation of PSM tables. So, it is expected to have the same PSM tables. Also, that’s a very good question. For those with missing precursor intensities, intensities are imputed using the global minimum MS1 intensity. Thanks for the suggestion. We will discuss if it is necessary to report the global minimum MS1 intensity, instead of NA, for better interpretation of the reports.

Best,

Huiyin

enonimos @.***> 於 2022年6月16日 上午2:48 寫道：



Hi Huiyin, Yes that's correct, I sent the PSM tables for v18 because the issue I saw isn't one that appeared after the update, it occurs in both v17 and v18 in the same way. The difference in FragPipe workflow is only in isobaric tab, "Use MS1 Intensity" was either enabled (MS1_Int file) or disabled (MS2_Int file).

In the TMT reports, the experimental channels have the same number of quantitative values independent of the MS1 intensity option, and as you said, the ratios are the same, while the abundances only differ systematically I assume due to different scaling fator. This all makes sense to me.

So the issue I raised only affects the RefInt/Pool channel column. If I understand what your saying with TMT-I filtering, there is the original reference channel coming from MS2, then if Use MS1 Intensity is enabled, the MS2 reference values are overwritten by the new MS1 values. However, if the MS1 intensity value is 0, then an NA will be inserted instead, and that's why I have more missing values for those columns.

If this is correct, then I have a follow-up question. How are the ratios converted to abundances for those case where MS1 intensity calculation fails? Does it default back to the MS2 intensity of the reference channel? If so, then shouldn't those values be listed in the report instead of NA?

Thank you in advance for the clarification.

Todd

— Reply to this email directly, view it on GitHub https://github.com/Nesvilab/FragPipe/issues/714#issuecomment-1156811466, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALAWWAY5NIAUUKRN4VYICYTVPIQP5ANCNFSM5YPK635Q . You are receiving this because you were mentioned.Message ID: @.***>

[enonimos]

Hi Huiyin,

I see, the global MS1 minimum makes sense. I agree if this value were to be written into the RefInt columns, then that would avoid confusion.

Thanks, Todd

[huiyinc]

Hi Todd,

Thanks for your valuable feedback and suggestions.

Huiyin

enonimos @.***> 於 2022年6月18日 上午12:51 寫道：

 Hi Huiyin,

I see, the global MS1 minimum makes sense. I agree if this value were to be written into the RefInt columns, then that would avoid confusion.

Thanks, Todd

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.