Reading the title, I was excited: It’s open access at Nature: Here’s the link and full citation:
Wu, Quan et al. 2015. «Suberoylanilide Hydroxamic Acid Treatment Reveals Crosstalks among Proteome, Ubiquitylome and Acetylome in Non-Small Cell Lung Cancer A549 Cell Line.» Scientific Reports 5: 9520. http://www.nature.com/doifinder/10.1038/srep09520.
Sure it’s only studying one cell line, but it’s a start. Integrating different post-translational modifications in proteomics data is super cool. By, using the term ‘crosstalk’ I was almost giddy with excitement. So, why the disparaging blog title?
First off, the text is hard to read and clearly not from native English speakers. That normally doesn’t bother me, but in this case, the prose is just super boring but and really, really hard to understand. For example:
“As it is well known that proteins undergo ubiquitination will be degraded by the proteasome due to the biological function of ubiquitination, we proposed previously that SAHA treatment mediated ubiquitination pathway probably is an unrevealed mechanism to regulate the proteome in A549 cells.”
Augh, WHAT? well known… proposed previously… probably is an unrevealed….. ???
“We found that SAHA treatment could largely change protein ubiquitination level both increase and decrease.”
I get what you’re saying here… but….. how about this version:
“SAHA treatment increases and decreases protein ubiquitination.”
Several times throughout the manuscript, the authors restate results and conclusions from their earlier work. I know we all cite our own work and often projects are serial and depend on previous results, but in this paper, I was asking myself, is this data here, or in another manuscript.
Speaking of data, where is the data? As of yet, no data relating to this study or their previous data has been deposited in Proteome Exchange or PRIDE or ProteomicsDB.
The experimental conditions of this work are great. The mass spectrometry instrumentation, the liquid phase chromatography conditions, the pan-antibody enrichment are all state of the art and are obviously being carried-out with expertise. The bioinformatics analyses are also strong, but nothing, seems novel. Nothing. Doing a bunch of heatmaps for gene set enrichment is not novel. And speaking of the heatmaps, both in the supplementary and in the primary manuscript, I needed to zoom in 4 times to read the categories. In print, they’d be difficult, if not impossible to read.
On the plus side, it was nice to see annotated mass spectra with and without a post-translational modification. But again, there were some ‘hair-ball’ networks that I could not read even when zoomed. Further, the legend and/or caption in Figure 5 (c, d, e, and f) doesn't include what the ‘red’ color represents. I will assume red indicates that the protein was found to be both ubiquitinated and acetylated.
I did appreciate Figure 3 A and B. These are what I think are the meat of the work. These figures represent the ‘cross-talk’ the authors are trying to convince us of. But, here’s the big problem. These figures show correlation, not crosstalk. And, weak correlation at that, -0.53 and -0.46 for Pearson’s and Spearman, respectively (oh, Pearson should be capitalized). Yes, Figure 3A definitely shows that as proteins have increasing levels of ubiquitination, their global quantification decreases. That’s the proven function of ubiquitination, right? Why is that a main conclusion from this work?
Figure 3B shows the relationship of proteins that contain both acetylation and ubiquitination on the same lysine site. There does seem to be a positive relationship, but there are only 43 proteins. And still, this is just a correlation, there’s nothing in the data that show direct crosstalk.
The other big conclusion presented in the paper, histones contain dense sites of acetylation that are increased by SAHA treatment. Yep, isn’t that the function of a HDAC inhibitor, to inhibit deacetylation? And here’s the pearl of this data that should have been the real focus, why are there any histone sites that show a decrease in acetylation after SAHA treatment? Was there an experimental error, or have you unwittingly uncovered something truly novel. And, yes, of course the top Gene Ontology categories enriched after SAHA treatment in the acetylome involve DNA replication, transcription factors, and transcription.
SO, both primary conclusions from the work are what I would consider expected. I realize that even if something is expected it must be proven empirically, but every conclusion has been known and proven, albeit on a smaller, non-’omics’ manner. Then, to use the term cross talk in the context of post-translational modifications was the offensive super stretch of the manuscript. Then to add insult to injury, the data appear to be strong, up to date, comprehensive, but not available.
So, how’d this get into Nature? Anyone?