If you are hoping to attempt a Trinity assembly, requirements for RAM = .5 * X million read pairs. For instance, to assemble 40 million paired-end reads using Trinity, you'll need a minimum of 20Gb of RAM.
These instructions work with a standard Ubuntu 16.04 or 18.04 machine available on AWS. Similar instructions should work for people on their own workstations, especially if you have sudo privileges, or even if you don't.
This is typically necessary only when starting from a fresh machine.
sudo apt-get update && sudo apt-get -y upgrade && sudo apt-get -y install build-essential git libxml2-dev libz-dev sudo nano wget curl
Packages are installed mostly via conda - at the end of the make process, you will have an orp conda environment, that will contain everything you need for assembly. Make sure to type source ~/.profile and the end of the make process/
git clone https://github.com/macmanes-lab/Oyster_River_Protocol.git
cd Oyster_River_Protocol
make
source ~/.profile
### Make sure to add the items to your profile file, as needed.
**Make sure to ```source``` the profile file after, to make sure everything is loaded.**
You should just have to change the user path info (the stuff before /Oyster_River_Protocol/...). A simple way to do this is via sed.
sed -i "s_ubuntu_$(whoami)_g" $HOME/Oyster_River_Protocol/software/config.ini
You may want to install additional BUSCO databases - the Euk. database is installed and used by default.
###Download databases
cd $HOME/Oyster_River_Protocol/busco_dbs
# Eukaryota
wget http://busco.ezlab.org/v2/datasets/fungi_odb9.tar.gz
wget http://busco.ezlab.org/v2/datasets/metazoa_odb9.tar.gz
wget http://busco.ezlab.org/v2/datasets/nematoda_odb9.tar.gz
wget http://busco.ezlab.org/v2/datasets/arthropoda_odb9.tar.gz
wget http://busco.ezlab.org/v2/datasets/insecta_odb9.tar.gz
wget http://busco.ezlab.org/v2/datasets/vertebrata_odb9.tar.gz
wget http://busco.ezlab.org/v2/datasets/tetrapoda_odb9.tar.gz
wget http://busco.ezlab.org/v2/datasets/aves_odb9.tar.gz
wget http://busco.ezlab.org/v2/datasets/mammalia_odb9.tar.gz
tar -zxf ..
cd
This is a very small data set that should assemble ~30 transcripts. It will finished in a few minutes or less using desktop-sized computer. The BUSCO numbers you get at the end are bad, for obvious reasons. If this finishes without error, you're good to move on to a 'real' assembly!! Good luck, and ping me on Gitter if issues!
You must activate the orp conda environment, that make made for you.
You must use the full PATH to the oyster.mk script for it to work
cd $HOME/Oyster_River_Protocol/sampledata
source activate orp
#note use of full PATH (your PATH to oyster.mk might be different)
$HOME/Oyster_River_Protocol/oyster.mk \
STRAND=RF \
TPM_FILT=0.2 \
MEM=15 \
CPU=8 \
READ1=test.1.fq.gz \
READ2=test.2.fq.gz \
RUNOUT=test
At the end of the successful run, you should see some text that looks like this. Your numbers will be different, but should similar. Assembly is not deterministic.
14| #
13| #
12| #
11| #
10| #
9| #
8| #
7| ##
6| ##
5| ##
4| ##
3| ##
2| ##
1| ## #
-----------
------------------------
| Summary |
------------------------
| observations: 20 |
| min value: -1.000000 |
| mean : -0.987400 |
| max value: -0.935000 |
------------------------
***** See the following link for interpretation *****
***** https://oyster-river-protocol.readthedocs.io/en/latest/strandexamine.html *****
***** QUALITY REPORT FOR: test using the ORP version 2.2.0 ****
***** THE ASSEMBLY CAN BE FOUND HERE: /root/ORP/sampledata/assemblies/test.ORP.fasta ****
***** BUSCO SCORE ~~~~~> C:0.0%[S:0.0%,D:0.0%],F:0.3%,M:99.7%,n:303
***** TRANSRATE SCORE ~~~~~> 0.37518
***** TRANSRATE OPTIMAL SCORE ~~~~~> 0.56393
***** UNIQUE GENES ORP ~~~~~> 39
***** UNIQUE GENES TRINITY ~~~~~> 31
***** UNIQUE GENES SPADES55 ~~~~~> 22
***** UNIQUE GENES SPADES75 ~~~~~> 23
***** UNIQUE GENES TRANSABYSS ~~~~~> 35