Tragedy of the Day - New 'Revealing' Human Transcriptome Paper at Biorxiv
Someone forwarded me the link to a new human transcriptome paper, and as an author of the first genome-wide human transcriptome paper exactly 10 years back, I decided to check how far the field advanced over the decade. What I found is quite shocking and it is abundantly clear that these ENCODE-clowns have not learned any lesson through their public humiliation.
The paper examines the transcriptomes of ~400 female twin pairs (~800 individuals), throws all kinds of statistical analysis tools and reports absolutely nothing interesting. No new pathways, mechanisms, nothing. Based on that summary, I thought the title would be ‘comparative analysis of transcriptomes of 800 UK twins shows nothing’. Instead the authors chose ‘transcriptome sequencing reveals widespread gene-gene and gene-environment interactions’. Oh, please !!
BS starts right from the abstract. The last statement of the abstract says - ‘We uncover large GxG and GxE effects on gene expression and likely complex phenotypes that currently remain elusive’. We pondered about the later part of the sentence and could find absolutely no meaning whatsoever. Do the authors mean ‘we uncover complex phenotypes that currently remain elusive’? How can you uncover something that you do not uncover anywhere in the paper? The best explanation is that the authors wanted to include the words ‘complex phenotype’, because that is what gets them money these days. Otherwise, they could as well say ‘we uncover likely the hand of God’ and be as speculative.
The main text starts with -
Gene expression is a cellular phenotype that informs about the functional state of the cell. Gene expression by itself is a complex trait that depends on genetic and environmental causes.
In plain English, they are saying ‘gene expression’ is now the new definition of function/phenotype, and with that broadened definition of phenotype, here comes our speculative paper. ENCODE 3 is reborn !! That is far from the original intention of such research projects, where they were supposed to explain real phenotypes, not redefine phenotype. However, as Sydney Brenner explained, the inverse problem of physiology cannot be solved for mathematical reasons. Hence comes the redefinition and hiding behind ‘complex phenotype’ metaphors.
A high point of the paper is to go through a long list of genes, find one or two well known members through cherry-picking process and then assign some kind of meaningless biological story. The authors picked Epstein-Barr virus induced 3 gene EBI2.
One of the top hits in LCLs is the Epstein-Barr virus induced 3 gene (EBI3) (Figure 3). That means that ASE at the EBI3 gene depends on the interaction of cis genetic variants and an environmental factor that, likely, in this case is the transformation process of the B cells with EBV.
By now, we have learned that all GWAS-type papers end with the obligatory statement that the authors could not uncover something, because their sample size was small. That later statement then becomes the first sentence in the follow-up grant application asking for more money to do the study with 4000 twins. This paper is no exception.
We found an example of GxE interaction on gene expression that has been widely described in the literature (Adiponectin) supporting the validity of our approach. However, the limitation on power due to the sample size prevented us to discover specific associations in most other cases.
Not that lack of funding was any obstacle for their work.
Acknowledgements: This work has been funded by the EU FP7 grant EuroBATS (No. 259749) which also supports AAB, AB, MND, AV, TDS. AAB is also supported by a grant from the South-Eastern Norway Health Authority, (No. 2011060). RD is supported by the Wellcome Trust (No. 098051). The Louis-Jeantet Foundation, Swiss National Science Foundation, Eurpean Research Council and the NIH-NIMH GTEx grant supports ETD. TS is an NIHR senior Investigator and holder of an ERC Advanced Principal Investigator award. JBR and HZ are supported by the Canadian Institutes of Health Research, Fonds de Recherche Sante du Quebec, and the Quebec Consortium for Drug Discovery. Most computations were performed at the Vital-IT (http://www.vital-it.ch) Center for high-performance computing of the SIB Swiss Institute of Bioinformatics. The TwinsUK study was funded by the Wellcome Trust; EC FP7(2007-2013) and National Institute of Health Research (NIHR). SNP Genotyping was performed by The Wellcome Trust Sanger Institute and National Eye Institute via NIH/CIDR.
In summary, this paper is so meaningless that we suspect it will be channeled into Nature through Magdalena Skipper.
