I have a large-ish dataset of the relative expression of about 8,000 proteins - I want to see what metabolic pathways change in this dataset, and how. I've been advised MFA is the way to go...can anyone point me in the right direction i.e. analysing SILAC data using MFA? Thanks!
Hi Vikram,
I am not sure if I understand your question. Metabolic flux analysis requires data about metabolites - I presume that your SILAC experiment contains data about proteins. If you really want to draw conclusions about flux, you probably also need several time points, rather than just data about the relative expression in two states.
I would suggest that you explore your dataset with a tool such as DAVID (http://david.abcc.ncifcrf.gov/) or Perseus (http://www.perseus-framework.org/) to see, if the proteins that are changing in your study are enriched in particular pathways (looking up databases such as REACTOME or using GO-tools).
Be aware that looking at 8000 proteins in what is probably only a low number of (biological? technical?) replicates poses also a significant statistical problem. You will need to ask yourself all the time the question: "what is the probability that I have obtained this result based on purely random fluctuation of the data?"
Just to add to the great answer above, I think you might be a bit confused in regards to flux analysis. Typically MFA is done on metabolites, normally by tracing relative C13 (via C13 glucose) incorporation in various metabolites (typically amino acids) gives you an idea of flux of consumption and production of metabolites. In regards to proteomics data, if you're looking for which proteins are upregulated/downregulated in your data set I would recommend reading up some large scale proteomics studies to get an idea of the software and statistical means used to achieve this. Like mentioned above, you will need replicates, and need to perform correlation analysis (probably some global normalization) between the replicates to determine if you can even use them.
Unless of course if you're really looking to do flux analysis via peptides, then there is a recent clever paper from my previous in which they do this:
http://www.ncbi.nlm.nih.gov/pubmed/25188426
Good luck!
Cheers for that guys - in response to your questions: I have SILAC data from 3 time points, from 3 biological replicates. The 8,000 proteins are across all 3 replicates in all time points and I've done extensive statistics mainly using R to narrow down my list. I've also done all the standard analyses - ORA (DAVID, ReviGo etc etc), GSEA...
Tanveer - ta for the paper; that's exactly what I was looking for! I was hoping to measure a deviation of sorts from a theoretical metabolic steady-state using my peptide/protein quantification - the paper seems perfect! I'll look through it just now.
Thanks again guys!
Actually I think what Vikram meant also was that he had 8000 proteins or gene expressions he would like to plug that data into a tool that would model how the metabolic route and flux of the pathway has changed. For example. If certain proteins in the glycolysis pathway are downregulated what happens to the route of glycolysis. Does the pathway shift to the pentose phosphate pathway etc. Are there any online tools that one can input quantitative proteomics data and determine how glycolysis is affected?
- Badges
- Science topic