Genetic Basis of Transcriptome Diversity in Drosophila melanogaster

This preprint from Trudy Mackay‘s group at NC State uses a recombinant inbred series of flies known as the Drosophila Genome Reference Panel (DGRP) to evaluate how genetic variation affects transcription.  To do this they extracted RNA from whole flies, pooled from males and females of each of the 192 recombinant inbred fly lines.  They fit the transcriptome data for each gene to a linear mixed effects model that included gender as the dependent variable and the RI line as a random effect.  They then calculated broad (based on the random effects term) and narrow sense (by incorporating a genetic covariance matrix) heritability for each transcript.

They found several interesting things.  First, with respect to sex they found that of the 18,140 transcripts tested, 7626 of these genes showed significant variability across the panel.  The broad sense heritability of genes ranged from 0.034-0.946 across the panel.

They then incorporated additive genetic variance into their model to compare narrow sense and broad sense heritability, and surprisingly found that this simple additive model explained most of the broad sense heritability (for a good explanation of the difference between the two see here, here and here).  What this suggests is that in their model, additive genetic effects explains most of the phenotypic variance of the transcripts.  They also modelled the within-transcript variance in expression, and found that unlike mean expression, the variance in expression was often modified by trans-eQTLs.  Supporting this, they found that nearly 50% of genes had a nearby cis-eQTL, mostly clustered around the 5′ transcriptional start site.  This is consistent with the idea that genetic drivers of transcriptional variation are near the upstream transcriptional regulatory modules of a gene.  The most surprising thing they saw (to me at least) was that most genetic expression variance was sex-specific, with only 185 eQTLs affecting expression in both sexes, out of 941 (male) and 1139 (female) cis-eQTL’s identified.  This is much higher than the rates identified by others in humans (12-17% in this paper).  Certainly this suggests that incorporating sex as a factor in GWA studies is extremely important.

In sum, this paper along with the previous work by Leonid Kruglyak’s group showing most heritability can be explained by additive genetic effects (see here in yeast) adds an important element to our understanding of how genetic variation can affect heritability.

Hat tip to Haldane’s Sieve who mentioned the paper here

Wen Huang, Mary Anna Carbone, Michael Magwire, Jason Peiffer, Richard Lyman, Eric Stone, Robert Anholt, & Trudy Mackay (2015). Genetic Basis of Transcriptome Diversity in Drosophila melanogaster bioRxiv : 10.1101/018325


The association between sleep duration and obesity amongst Canadians aged 12 and over

In this paper, published on PeerJ Preprints, the authors describe the relationship between sleep, obesity and hormones in adolescent Canadians.  The authors used data from the 2012 Canadian Community Health Survey (CCHS) to evaluate sleep duration and body mass index .  The authors looked both at adolescents (12-17) and adults (18+).  In both cases, obese and overweight participants slept less, with a negative correlation between sleep and obesity with correlation co-efficients of -0.031 and -0.056 for adolescents and adults respectively.  This is consistent with other studies that have also shown that sleep deficits either lead to, or correlated with increased obesity risk.

  • I am curious if the lack of sleep corresponded to changes in activity level or food intake, if that data is in the survey, or whether sleep deprivation is an independent predictor of obesity.
  • A plot showing the correlation would be helpful.  The analysis presumes a linear relationship between sleep and obesity, but if it is a nonlinear relationship, a more powerful model could be used.
  • Was the relationship similar between males and females?
  • The p-values are missing in Table 3
  • It would be helpful if the introduction and discussion had some more details of some of the other proposed causes of obesity, in addition to sleep disruption.

As a side note, this project came out of a undergraduate research project, so congratulations are due to both Maha Temkit and her advisor, Dr. Sanni Yaya for publishing an interesting UROP project!

Temkit, Maha, & Yaya, Sanni (2015). The association between sleep duration and obesity amongst Canadians aged 12 and over PeerJ Preprints : 10.7287/peerj.preprints.962v1


Preprints are papers that are shared, discussed or available prior to peer review publication.  This may sound new, or odd but in reality this happens in closed doors all the time.  People write papers, share them with trusted colleagues and then submit them to journals for peer review.  This would work fine, but for many reasons, papers often take a very long time to get published.  This means that we can’t build on, or apply that work unless we are one of the trusted colleagues, or until it is published.  This concept has been described in detail in Science, in a PLOS One article, and many other places.

The Solution?

Other fields, physics and math especially, work differently.  They post paper before, or simultaneous to peer review on preprint servers (mainly arXiv), for discussion and review by a wider audience than just the small number of peer reviewers chosen by the journal.  Biology has been historically resistant to this idea, but recent services such as bioRxiv and Peer J Preprints have provided a platform for posting preprints.

The Problem with the Solution…

My sense is, right now there is insufficient engagement, commenting and advertisement of preprints, especially in the areas that I am interested in.  My group has started posting our manuscripts to bioRxiv (I discussed that rationale here) for comments, and has received much more detailed comments and suggestions in detailed peer-review than anything that came from these preprints.  I suspect that my experience is not atypical.  Even if preprints are being read, and considered, I think that right now there is not sufficient public discussion on these preprints, and if they are going to be useful, they need to be discussed.

What I Hope to Accomplish

I plan to use this blog as a forum to read and make some comments on pre-prints that I come accross that  are in my area.  My goal is to advertise them better, provide some (hopefully) useful feedback to the authors and to force myself to consider this route of discourse.  Basically, I am stealing this idea from Haldane’s Sieve, a blog that discusses preprints in the area of population and evolutionary genetics.  Also, if you would like to contribute, and have something to say about a pre-print, let me know.

Desjardins-Proulx, P., White, E., Adamson, J., Ram, K., Poisot, T., & Gravel, D. (2013). The Case for Open Preprints in Biology PLoS Biology, 11 (5) DOI: 10.1371/journal.pbio.1001563