glass ceiling - cant find any more proteins/peptides

Search algorithms, post-searching processing, quantitation software, etc. Share and discuss software here.
neil777
Angiotensin Member
Angiotensin Member
Posts: 27
Joined: Fri Jan 20, 2012 9:53 am

glass ceiling - cant find any more proteins/peptides

Postby neil777 » Thu Sep 05, 2013 5:39 am

Hi I have been working on analysis of cell lysates etc and want to mine down to lesser abundant proteins/peptides.
my standard workflow is
fasp like lysis, followed by digestion with trypsin,
i then fractionate my sample into 24 using a mixed mode wax1 column
i then run each sample separately (on a Bruker Impact ESI TOF) on a 1 hour gradient
The best data i have got for combining all 24 runs together is about 3000 proteins and 20,000 peptides

I have tried to 'squeeze' as much data out of the mass spec as possible, for a typical run i might find as many as 700/800 proteins per fraction for a 2 hr gradient which is about as good as it gets.

I see that some people can get alot more proteins/peptides than this and wondered if anyone has any tips.
(I cannot afford to buy a thermo instrument !)
best wishes
neil

Infinity
E. Coli Lysate Member
E. Coli Lysate Member
Posts: 107
Joined: Wed Dec 21, 2011 8:22 pm

Postby Infinity » Thu Sep 05, 2013 6:57 am

If you can not use another instrument you might improve number of identified proteins by separating your sample on SDS-PAGE and then performing in-gel digestions of the bands (I know it is more time consuming and labor intense).

tsbatth
Ubiquitin Member
Ubiquitin Member
Posts: 58
Joined: Thu Jul 12, 2012 7:26 pm

Postby tsbatth » Thu Sep 05, 2013 12:07 pm

Have you checked for overlap of your peptides, proteins across fractions ? You might be getting lots of overlap. Fractionate your samples using high pH reverse phase if possible. Concatenate it to around 10-15 fractions and run 2hr gradients on each. See if that helps.

botte
Phosphoserine Member
Phosphoserine Member
Posts: 20
Joined: Mon Apr 29, 2013 11:47 pm

Postby botte » Thu Sep 12, 2013 5:47 am

What kind of chromatography setup are you using coupled to your MS? Especially for longer gradients, use of longer columns does make a difference. Apart from that I would second Infinity's suggestion - we get good results from Sd
SDS-PAGE with Coomassie staining, slicing 23x irrespecti e of staining an running 1h gradients per slice. Relying just on the FASP might render some subpopulations poorly accessible. Imho FASP is mechanistically not well understood.

Omics
Phosphoserine Member
Phosphoserine Member
Posts: 14
Joined: Sat Oct 06, 2012 10:34 am

Postby Omics » Thu Sep 12, 2013 7:49 am

The sensitivity of mass spectrometry is the most important factor for increasing peptide and protein IDs. The concatenation strategy is helpful in identifying more peptides, but not more proteins, as it may be limited by the sensitivty of your mass spectrometry. If you have to use the same instrument, the use of multiple fractionation methods, such as ERLIC and HP-RP, may be a reasonable way to increase protein and peptide IDs. Please refer to this article: Hao et al. J Proteomics. 2013 Apr 26;82:254-62.

tsbatth
Ubiquitin Member
Ubiquitin Member
Posts: 58
Joined: Thu Jul 12, 2012 7:26 pm

Postby tsbatth » Thu Sep 12, 2013 8:55 am

Omics wrote:The sensitivity of mass spectrometry is the most important factor for increasing peptide and protein IDs. The concatenation strategy is helpful in identifying more peptides, but not more proteins, as it may be limited by the sensitivty of your mass spectrometry. If you have to use the same instrument, the use of multiple fractionation methods, such as ERLIC and HP-RP, may be a reasonable way to increase protein and peptide IDs. Please refer to this article: Hao et al. J Proteomics. 2013 Apr 26;82:254-62.


Increased identification of more peptides should in theory lead to higher number of proteins hits. In that study you referred to, there wasn't much difference between the two (concatenated ERLIC and HP-RP) in terms of peptide #'s and proteins, I'm not sure how performing both will increase the number of proteins or peptides. In fact, the difference that are presented are well within the margin of error, they use technical duplicates for each (not sure how they got away with error bars from those). I think best way would be to use SCX with high pH or SDS, since the fractionation techniques are inherently different and might yield different populations of peptides.

Omics
Phosphoserine Member
Phosphoserine Member
Posts: 14
Joined: Sat Oct 06, 2012 10:34 am

Postby Omics » Thu Sep 12, 2013 8:08 pm

tsbatth wrote:Increased identification of more peptides should in theory lead to higher number of proteins hits. In that study you referred to, there wasn't much difference between the two (concatenated ERLIC and HP-RP) in terms of peptide #'s and proteins, I'm not sure how performing both will increase the number of proteins or peptides. In fact, the difference that are presented are well within the margin of error, they use technical duplicates for each (not sure how they got away with error bars from those). I think best way would be to use SCX with high pH or SDS, since the fractionation techniques are inherently different and might yield different populations of peptides.


Even if the total number of protein and peptide IDs is similar or identical for different fractionation methods, different fractionation methods yield different populations of peptides, that is why the use of multiple fractionation methods gets more protein and peptide IDs. It is one of the conlusions of the article I mentioned. As ERLIC, SCX and HP-RP separate peptides based on different properties, the use of any two of them will result in the increase of peptide and protein IDs. However, it is reported that SCX is not comparable with either ERLIC or HP-RP in peptide and protein IDs, that is why I suggest the use of both ERLIC and Hp-RP.

neil777
Angiotensin Member
Angiotensin Member
Posts: 27
Joined: Fri Jan 20, 2012 9:53 am

Postby neil777 » Fri Sep 13, 2013 7:46 am

Hi,
i am using dimethyl labelling of peptides, therefore i cant really separate proteins on gels as i need to mix the labelled peptides together and then separate them. out of interest how many proteins might you expect to identify from running a lysate on SDS page
best wishes
neil

neil777
Angiotensin Member
Angiotensin Member
Posts: 27
Joined: Fri Jan 20, 2012 9:53 am

Postby neil777 » Fri Sep 13, 2013 7:50 am

To omics,
i think that the sensitivity of the instrument might be the limiting factor
i guess i could load a higher concentration of peptides but have to be careful not to 'overload ' the C18 trap and C18 column (25cm)
i wonder how much total peptide (in ugs) other people tend to load on their systems
cheers
neil

Omics
Phosphoserine Member
Phosphoserine Member
Posts: 14
Joined: Sat Oct 06, 2012 10:34 am

Postby Omics » Sat Sep 14, 2013 12:12 pm

Generally, for 15cmX75um column, the maximum load amount is 2 ug. Overloading has a negative effect. Based on my test, the resolution of SDS-PAGE is poor, and it is much worse than any peptide fraction methods. If anyone else also did such a comparison, I am glad to have a discussion about it.

Another way to improve protein and peptide ID is to improve your data analysis methods. For example, the machine learning methods always produce a better result than the FDR cutoff methods based only on peptide score.

tsbatth
Ubiquitin Member
Ubiquitin Member
Posts: 58
Joined: Thu Jul 12, 2012 7:26 pm

Postby tsbatth » Thu Sep 19, 2013 9:25 am

tsbatth wrote:Increased identification of more peptides should in theory lead to higher number of proteins hits. In that study you referred to, there wasn't much difference between the two (concatenated ERLIC and HP-RP) in terms of peptide #'s and proteins, I'm not sure how performing both will increase the number of proteins or peptides. In fact, the difference that are presented are well within the margin of error, they use technical duplicates for each (not sure how they got away with error bars from those). I think best way would be to use SCX with high pH or SDS, since the fractionation techniques are inherently different and might yield different populations of peptides.


Whops what I meant was, I was going off the fact not that the numbers were similar (they were), but the identified peptides/proteins from ERLIC/high pH were more or less the same.

Niel, you could try running 4 hour gradients on a 50cm column if you have a nano/uplc, that could potentially help.


Return to “Bioinformatics”

Who is online

Users browsing this forum: No registered users and 1 guest