Recently i started a proteomics of blood plasma (100 ms/ms files .raw) against fungal fasta (from uniprot) using MaxQuant in linux. I started with 6 .raw then 30 and finally all 100 files. In 6 files i was able to detect about 97 proteins, in 30 it was 127 but in case of 100 files where it should be much higher, count was reduced to 67 and most of the detected entries were of contaminants (100 in 167). Also,null intensity count is over 50%. So, I am stuck at this point. Beside this, i also performed the extraction of all these 100 files in the batches of 10. In this case total 666 proteins were detected. I don't know if i can trust this method or not. But, why am i not able to get this in one single go? Below are some parameters and system specifications i used for the analysis.
Parameters
Fixed modifications: carbamidomethyl (c)
enzyme: Trypsin/P
Variable modifications: Oxidation(M), Acetyl (Protein N-term)
3 groups, LFQ
Peptide, protein and site FDR: 0.05, 0.05 and 0.05
System
RAM-128GB
CPU- intel i9 11th generation, 16 cores
Working memory- 2TB SSD,
NVDIA Geforce RTX 3080, 16 gb
Please do the needful.
Hi Rohit Kumar Nadda,
did you enabled the "match between runs" function?
Best,
Murat
Hi Murat Eravci , thank you for response. Yes, match between runs function is on. I set the parameters under this as below:
Match time window(min.): 0.7
match ion mobility window 0.05
Alignment time window(min.): 20
alignment ion mobility window: 1
match unidentified features: unchecked.
Hi again,
this is strange. in my experience (and also from a logical point of view - regarding the match between runs function) the number of identified proteins should be higher with the number of raw files increasing.
However there are two things that I want to ask. First, the number of proteins in your experiment are way too low. From blood plasma you should obtain (depending on your LC-MS System) a few thousand proteins instead of hundreds, even with one single raw file. I would recommend to troubleshoot on this issue first (i.e. sample prep, and performance of the MS platform used). Second, when doing 100 LC-MS runs after each other, the retention times from first to the last sample can be very different for this reason it might be good to increase the retention time window in the match between runs setup. If the runs are not performed in one session, maybe in multiple sessions even with different columns I would recommend to do the MQ analysis in the same way (session wise).
Best,
Murat
Hi Murat Eravci thank you for the response. In this study i have only used 3 to 4 of fungal species. So fasta is quite short (700-1000kb) and specific and that could be the reason for number of proteins. I am saying this because we have detected >1500 proteins with human fasta (same raw files). Can you please recommend a suitable "match time" and "alignment time window"? Because LC-MS runs for all 100 were performed in one session.
Regards,
Rohit
Hi Rohit,
I can not give any suitable match time and alignment time window since it is dependent on your LC-MS performance and the level of possible drift of retention times from the first to the last sample. maybe you can try to inclease the retention time window in steps of +30 seconds and look if you get more identifications.
In relation to your FASTA, I would suggest to take a combined FASTA from all available fungal species in close taxonomy to your species of interest to increase the database by entries of very homologue sequences from other very similar fungal species.
Best,
Murat.
- Badges
- Science topic
- Similar topics
- Biological Science
- Biochemistry
- Biomolecules
- Peptides