Populus tremula shows no evidence of sexual dimorphism

Keywords

FASTQ, GFF3, BAM, Populus-tremula, RNA-Seq, Pre-processing, QC, Alignment, Annotation, Expression-estimation, Differential-expression

Authors

• Nicolas Delhomme (@delhomme)
• Robinson K.
• Mähler N.
• Bastian Schiffthaler (@bastian)
• Oenskog J.
• Albrectsen B.
• Ingvarsson P. K.
• Hvidsten T. R.
• Jansson S.
• Street N. R.

Description

Background: Although the majority of plant species are co-sexual, being either monoecious or hermaphroditic, a significant number are dioecious, having separate male and female individuals. Evolutionary theory suggests that males and females may develop sexually dimorphic phenotypic and biochemical traits concordant with each sex having different optimal strategies of resource investment to maximise reproductive success and fitness. The establishment of such sexual dimorphism would result in changes in gene expression patterns in non-floral organs.

Results: We examined phenotypic, biochemical and leaf herbivory traits to reveal evidence of sexually dimorphic resource allocation strategies within a collection of sexually mature and immature Populus tremula (European aspen) trees. We additionally generated gene expression data from mature leaves of sexually mature wild trees using whole-genome oligo microarrays and RNA-Sequencing. Statistical analysis of the phenotypic traits revealed no evidence of sexual dimorphism or differential resource investment strategies between males and females. Similarly, single-gene differential expression and machine learning on combinations of genes identified no statistically robust evidence of expression level sexual dimorphism. Analysis of the RNA-Sequencing data did not identify any sex-specific Single Nucleotide Polymorphisms or genomic regions with contrasting SNP rates.

Conclusions: We found no evidence of sexual dimorphism at the phenotypic, biochemical or genomic scales in P. tremula either before or after development phase transition (i.e. sexual maturity). There was no evidence that P. tremula males and females have evolved differential resource investment strategies or that they exhibit contrasting resistance or tolerance to leaf herbivores.

RNA-Seq data: Total RNA preparations were sent to the Science for Life Laboratory (SciLifeLab, Stockholm, Sweden) for sequencing. Paired-end (2 × 100 bp) RNA-Seq data were generated using standard Illumina protocols and kits (TruSeq SBS KIT-HS v3, FC-401-3001; TruSeq PE Cluster Kit v3, PE-401-3001) and all sequencing was performed using the Illumina HiSeq 2000 platform. We generated data from 8 male individuals (five sampled in 2008 and three in 2010) and 9 female individuals (five sampled in 2008 and four in 2010). For sequencing, samples were recoded (from 1-17) with males and females randomised to avoid bias due to sample handling order. Samples were multiplexed by the addition of a unique barcode sequence and all samples were profiled on two lanes of the same flowcell with male and female samples and samples from 2008 and 2010 randomised between the two lanes.

Particular feature:

The statistical analysis is very comprehensive and very educational due to the presence of a strong confounding factor (the sampling year) that severely bias the analysis if ignored. It helps introduce trainees to confounding factors, blocking and the importance of a proper study design.
The alignments are performed against the closest relative species: Populus trichocarpa as the genome of Populus tremula is not available yet. The sequence difference between the 2 genomes within their gene space is about 1 SNP/indel every 100bp.

Properties

• RNA-Seq
• FASTQ files - subset of 1M reads were extracted for teaching purposes
• Designed for pre-processing, QC, alignment, annotation, expression-estimation and differential-expression. A very interesting feature of this dataset is the presence of a strong confounding factor (the sampling year) that severely bias the analysis if ignored.
• Populus-tremula
• 17 RNA-Seq samples, originally with an average 20M PE reads per sample, downsampled to about 1M reads per sample.
• Altogether, the computationally intensive tasks may require up to about 2h. The corresponding teaching material is meant to be run over 1 and a 1/2 day.

Protocol:

The data subset has been manually adapted to retain the characteristics of the original dataset. Please contact @bastian for more details about this. Reproducing the data subsetting should not be necessary as the subset is available from the FTP below and the data public; but please contact @delhomme and @bastian if you disagree :-)

Top | Keywords | Authors | Description | Dataset links | Tutorials

Dataset-links

Data subset being provided

!!These links are not active yet!!

Data: an archive of 17 FASTQ files downsampled to roughly 1M PE reads each. It also contains a file describing the samples (e.g. year of sampling, sex, etc.) and the GFF3 annotation, the FASTA file and the STAR aligner indexes of the Populus trichocarpa genome.
Package: an R package to be installed prior to the course that contains all the tutorial and hands-on material.
Docker: a Docker (a self contained environment) based on the Bioconductor NGS Docker that can be used to setup the course machines (physical or in the cloud).

Original publication

• BMC Plant Biology, 2014

Original dataset

• ENA ERP002471

Top | Keywords | Authors | Description | Dataset links | Tutorials

Tutorials

### EMBO October 2014
* EMBO October 2014 - Course
* EMBO October 2014 - Content
* EMBO October 2014 - Tutorial
* EMBO October 2014 - Introduction
* EMBO October 2014 - QC
* EMBO October 2014 - Alignment
* EMBO October 2014 - Annotation
* EMBO October 2014 - Expression-estimate
* EMBO October 2014 - Differential-expression

Top | Keywords | Authors | Description | Dataset links | Tutorials