A big difference between Philosopher and Proteome Discoverer

[Bin-cc]

Hi,

I used to search proteomics raw data by Proteome Discoverer but now I need to process large-scale samples so I would like to use the pipline of Philosopher. It's really powerful and convenient. However I found that the results of Philosopher and of Proteome Discoverer from same test raw data (adding DMSO to cell lysate and searching by DDA) could not be completely coincident, about 50% overlap ratio.

So I want to enquiry that

1. Whether my parameters are ringht or not? (I attach the result of PD and Philosopher I have done and the Philosopher with the method I refer to this). If you need the raw data, please let me know.
2. Have you ever done the similar comparation between the search engine of Sequest in PD and Comet in Philosopher?

Best regards

PD_result, Philosopher_result.xlsx

[prvst]

Hi @Bin-cc. I handle PD and FragPipe/Philosopher results routinely, and based on my experience, I observe that in most of the cases we have a large overlap, with MSFragger-based runs resulting in more IDs. Some times the overlap is almost perfect, sometimes we get 30% more with Fragger. I depends a lot on the run settings, and the type of the experiment. It's hard for me to judge your case solely on your report, but 50% seems a little "too different". Perhaps you need to adjust some of the search parameters. You can do a lot of customization in FragPipe/Philosopher, so my suggestion to you is to start with the FragPipe GUI program, and using one of the pre-configured workflows.

[Bin-cc]

Thank you for your replying. I will try to adjust the search parameters.

[Bin-cc]

Hi,

I followed your advice and used the latest version of FragPipe, MSFragger and Philosopher with the workflow LFQ-MBR to search my data. And I also changed the parameters to make them as same as those I used in PD, but I still got the relatively low overlap ( about 66 %) between PD and MSFragger.

1. whether this result is right?
2. If this result isn't correct, could you help me to adjust the parameters?

Best regards

[prvst]

It's complicated to define what is "right" or "wrong", especially without knowing nothing about the experiment, the instrument, the methods, the protocols and the analysis. Without having access to the data and both programs you're running, all I can tell you is that, based on our (and others) comparisons, FragPipe performs better than PD. You might have a different type of data that for an unknown reason to me, performs better with PD, but again, without having knowledge about all the details, it is really hard for me to tell what is different and even more important, if what you're doing is "right". Regardless of what tool you're using, the general overlap between the most popular ones today, with a simple DDA LFQ run, would unlikely be 60%.

[Bin-cc]

Thank you for your patient answer. Maybe I need to review my method to find the solution. Best regards bin liao ***@***.*** 签名由 网易邮箱大师 定制 On 04/11/2022 21:18, Felipe da Veiga Leprevost wrote: It's complicated to define what is "right" or "wrong", especially without knowing nothing about the experiment, the instrument, the methods, the protocols and the analysis. Without having access to the data and bot programs you're running, all I can tell you is that, based on our (and others) comparisons, FragPipe performs better than PD. You might have a different type of data that for an unknown reason to me, performs better with PD, but again, without having knowledge about all the details, it is really hard for me to tell what is different and even more important, if what you're doing is "right". Regardless of what tool you're using, the general overlap between the most popular ones today, with a simple DDA LFQ run, would unlikely be 60% different. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[prvst]

Sorry for not be able to give you a conclusive answer. For now, I can suggest that you review your parameters one more time, including the PD workflows that can be very tricky to configure. If after reviewing everything, you still believe FragPipe is underperforming, then I urge you to reach out to us, including Alexey, via email.

[Bin-cc]

Thank you for your advice. I will check the workflow again of both the PD and FrigPipe. bin liao ***@***.*** 签名由 网易邮箱大师 定制 On 04/11/2022 22:13, Felipe da Veiga Leprevost wrote: Sorry for not be able to give you a conclusive answer. For now, I can suggest that you review your parameters one more time, including the PD workflows that can be very tricky to configure. If after reviewing everything, you still believe FragPipe is underperforming, then I urge you to reach out to us, including Alexey, via email. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>