Transcript Document
Making sense out of metabolite profiling data Scott Harding Feb. 19, 2014 Data interpretation: 1. So looking at the preliminary results of the metalab that I grouped into amino acids, sugars and alcohols, and the phenolics and other compounds my question then is “what the relationship of these data to each other? 2. What are the key points that we need to pay attention in terms of other compounds data in relation to biochemical processes and functions? 3. As guide for us, how do you want us to approach the metalab data? 4. When looking at different metabolites like the Amino Acids, I understand their importance for informing on carbon partitioning or protein synthesis but can you give some guidelines for what they could be telling us? For example, if one sample had a much larger amount of Alanine while another sample had a much larger amount of Threonine what sorts of things might this tell us about carbon and proteins? 5. With all the metabolites that are given in the raw files we go after a specific set of metabolites usually sugars, alcohols, and amino acids. As a general guideline to decrease the number of metabolites, what other groups are beneficial for us to go after? 6. Going with the question above, if your objecting is simply to find interesting patterns, what is the best way to go about finding these patterns? Which metabolites should be the most important ones to start with? What groups of metabolites should be included in this objective? Again, is there a set our lab focuses more on? Should we start by “eliminating” metabolites that have a low confidence? Are there a time point cutoffs where it is most likely just noise so that we focus on metabolites before/after those time points? 2) What are the different types of compounds from gs vs lcms and why is one set important biologically vs the other. I realize that you can get phenolics, but why would those be important to me biologically? 3) Is there any indication from metalab-type compounds what is going on at a large scale metabolic products like cellulose/lignin/bulkprotein? I realize that it's just measuring primary metabolic compounds, but is there any break down? I guess I'm wondering what limitations we have biologically drawing conclusions about large scale partitioning? Harding et al. 2014 Tree Physiology maltose serine glycine sucrose glucose Glucose-6-P fructose alanine glyceric sedoheptulose Ribose Erythrose Lyxose threose xylose quinic chlorogenic shikimic citric mucic saccharic gluconic a-ketoglutaric Pyruvic (lactic) succinic leucine valine malic asparagine oxaloacetic aspartic lysine threonine isoleucine Redox control mannose kaempferol catechin Benzoic… salicin catechol catechol-gluc proline Citrullene Ornithine Arginine fumaric caffeic paracoumaric tyrosine phenylalanine tryptophan glutamine glutamic CT populin salicortin ascorbic dehydroascorbic myo/alloinositol galactose galactinol raffinose Use regression when you want to evaluate whether a series of closely migrating peaks are just multiple peaks of one substance. • Sugar alcohols vs. acids vs. ketose/aldose • Sugar alcohols come from reduction of sugars while sugar acids come from oxidation of sugars. Not common in nature. • If you do a regression test between say ribitol, ribose or ribonic-acid, ribose… • If correlation is ‘good’, then… • If correlation not ‘good’ then… • Outcomes: sum or not to sum. putrescine degradation IV : putrescine + oxygen + H2O → ammonium + hydrogen peroxide + 4-aminobutanal Misleading/bad/false GC-MS peaks RT Name 5.8943 [8871] 2-hydroxypyridine [6.519] 6.2214 Tris(trimethylsilyl)hydroxylamine 6.4861 Pentasiloxane, dodecamethyl- 6.3667 OXALIC ACID 7.2747 [1004] phosphoric acid [9.966] 13.5218 [985] palmitic acid [18.846] 13.5329 SINAPIC ACID (1) 15.531 [5281] stearic acid [20.675] • The list will be updated regularly in the future • The yellow indicates two names for one peak, sometimes called palmitic sometimes called sinapic, both misleading since they are metabolite names, but these peaks are fatty acid garbage that just hangs up in the system and contaminates all of our samples. – Remember the RT values may shift depending on the column age