Assembling tall fescue transcriptomes with Trinity

This supplemental tutorial explains how assemble the RNA-seq reads for the six tall fescue samples from the end-to-end "From RNA-seq reads to a phylogenetic tree with RNA-clique" tutorial into transcriptomes using the Trinity assembler. This tutorial is a companion to the "Using RNA-clique with non-SPAdes data" tutorial, and since that tutorial provides download links for pre-assembled transcriptomes, this assembly tutorial should be considered optional.

Setup

Installing Trinity

Trinity provides an installation guide, but a different explanation of how to install Trinity via package managers is also provided here.

Note

Your system's package manager might not have the newest version of Trinity.

On Ubuntu, the package to install is trinityrnaseq. On macOS, Trinity is not available through Homebrew's core tap but is available through brewsci/bio as trinity.

UbuntumacOS

sudo apt install trinityrnaseq

brew tap brewsci/bio
brew install trinity

The newly installed Trinity program can be used via the Trinity command.

Installing GNU Parallel (Optional)

If you have not already, consider installing GNU Parallel. GNU Parallel will make it easier to run Trinity for multiple sets of data efficiently on multi-core machines.

Downloading RNA-seq reads

The process of downloading the tall fescue RNA-seq reads from the NCBI Sequence Read Archive is described in the Obtaining sequence data section of the end-to-end tutorial. That section also requires that some previous sections of the tutorial be completed. Specifically, it is expected that the Setup and Creating a directory for our work sections have been completed.

It will be assumed that the SRR*.fastq files downloaded in the end-to-end tutorial are in the $TUTORIAL_DIR for the remainder of this tutorial.

Assembling transcriptomes

First, change to the $TUTORIAL_DIR.

cd "$TUTORIAL_DIR"

We will run Trinity with multiple CPU cores and a memory limit. What specific settings you should use here depends on what hardware you have available on your machine.

CPU usage

The product of the number of Parallel jobs and the number of CPUs used by Trinity should not exceed the number of logical cores you have on your system, and you should prefer to use more Parallel jobs rather than more CPUs with Trinity. If you're not using Parallel, then the number of CPUs used by Trinity must not exceed the number of logical cores you have.

Memory usage

The product of the number of Parallel jobs and memory limit should not exceed the amount of RAM you have in your system. If you are not using Parallel, then the memory limit must not exceed the amount of RAM you have in your system.

The examples below are written for a machine with 32 cores and over 300 GB of RAM. You will need to change the parameters to fit your own system.

With Parallel

parallel -j 6 Trinity --seqType fq --max_memory 48G --single {} --CPU 2 \
         --output trinity_{/.} ::: *.fastq

Without Parallel

for f in *.fastq; do
    s="${f%%.*}";
    Trinity --seqType fq --max_memory 256G --single "$f" --CPU 32 \
            --output "trinity_$s";
done