I have tried to identify and enrich phospho-peptides by using TiO2 beads (control vs cisplatin treated Jurkat T-cells). Then I searched all the raw files using Proteome Discorverer 1.4 against Mascot database. All the output MSF files was analyzed by Scaffolds for label-free quantification.
The result I got mentioning the Mascot version used was Mascot 5 (which has less entries compared to normal swissprot (more than 20.000)), this is quite confusing to me.
I don't know what is Mascot 5 homo-sapien unknown version? It run on 16 databases, which is pretty strange. I have 16 samples. However, they should all run on the same database, right?
Furthermore, the peptide reports that I got from Scaffolds also mentioned some unknown proteins from the peptide #.... Which I never seen before.
So, I am just wondering if anyone have seen this problem before or have any suggestion how to analyse these label free quantification better?. Thank you very much.
Dear Trung,
It looks like you do not have one database for the searches you performed on Mascot 5. You have 16 databases, and they are all a bit different from each other. How did you set your parameters? It is best if you opt for Swissprot and under species, choose Human for each of the searches. If you save the parameters this way, each time you search Mascot, it should keep the same number of proteins/database to query your unknown dataset against.
The import into Scaffold requires THE SAME, EAXCT SWISSPROT database that was used for your Mascot search. If not, you will have ? marks. You can easily remedy this problem by going to "Experiment" on pulldown menu and "Apply New Database" activated. You need to make sure the database you have used for the Mascot search is in your Scaffold Database folder. If not, find Mascot SwissProt fasta file and under go back to Apply New Database, click on "Add Database", choose the Fasta file and Autoparse it while it is imported into the Scaffold Database. When it is in the Scaffold Database, just Apply it and you should be able to get rid of the ? marks under Molecular weight of the Proteins.
Best,
Hediye
Hi Trung,
first off some clarification, Mascot is not a database, it is a search algorithm that lets you search databases using MS data. In your case, you used Homo sapiens. The most recent version of Mascot is 2.5.1. In the screen shot, the database is named Mascot5_SwissProt_Homo sapiens. What is unusual is that you have the same database name, but the number of entries varies. To me, this suggests the database has changed.
And you are clearly getting matches that aren't in the database you've loaded into Scaffold. Hence the "unknown protein for peptide #.' My best guess is that you've searched against one database and loaded one, or more, similar but not identical databases when you put that data into Scaffold.
You need to ensure the database you search using PD/Mascot is identical to that loaded into Scaffold. I would suggest you check this and re-search is necessary.
Dear Hediye and Peter,
Thank you so much for your suggestion. I have tried to search all of my raw files again using Mascot algorithm but not Proteome Discorverer this time. Then I loaded all of the Dat files on Scaffolds using the same Swissprot database as the one I used on the Mascot search. And it solved my problems. I have used the same template parameters for all the files in Mascot search. I think it might be because of PD. I am not sure as I also use the same template for all the search in PD, might be the database version I used on PD was different from the one I used on Scaffolds.
Again thank you very much
Does anyone here know what could be the best approach to analyse label free quantification. I have tried to use Scaffolds, Maxquant-Perseus (LFQ approach), and Skyline. And since I am pretty new at this, they are quite complicated. If anyone here has some experiences in this, please share your suggestions. I am really appreciate all the help. Thank you so much.
Hi Trung,
I'm glad you worked it out. As for LFQ, I'm a big fan of Skyline, they have excellent tutorials on their website. Whilst it can seem a bit daunting at first, it's worth persevering as the software is very flexible. MaxQuant is popular and widely used. They also have a good support base through the Google group. Personally, I would think Scaffold would be the easiest implementation for a new user. As a paying customer you also have the company support if you run into problems.
Good luck,
Peter
- Badges
- Science topic
- Similar topics
- Biological Science
- Biochemistry
- Biomolecules
- Peptides