De Novo Transcriptome Assemblers – Oases, Trinity, etc. – III
Oases and Trinity Quick Start Guide
This guide is for those, who like to get started quickly with available transcriptome assemblers. So far, our experiences had been with Oases and Trinity only. If we become familiar with other transcriptome assemblers, this guide will be expanded.
Oases
Oases is the transcriptome arm of Velvet, a popular de Bruijn graph-based genome assembler.
Where do we download it from?
Please use following links - Velvet, Oases
How do we run it?
First compile and install Velvet and Oases. Then, run Velvet assembler on your reads just like you would do for genome assembly, but remember to add -read_trkg option. After Velvet completes, run oases.
Is there a guide with step-by-step instructions?
The instruction manuals at Velvet and Oases website are very helpful. Also, here is a nice guide with step-by-step instruction on how to run the programs. Advanced users are recommended to join Velvet and Oases mailing lists at this and this link for more complex questions.
Memory requirement
Oases is very memory intensive and requires even more RAM than the genome assembler Velvet. For a paired-end library of ~7M (100ntx2) reads, Oases used ~50-55Gb RAM.
Running Time
Velvet+Oases run much faster than Trinity, the other transcriptome assembler we cover here. Assembly time for a paired-end library of ~7M (100ntx2) reads is typically about 5-6 hours.
Trinity
Trinity is a stand-alone de Bruijn graph-based de novo transcriptome assembler from Broad institute. It is combination of three programs Inchworm (C++), Chrysalis (C++) and Butterfly (Java). You can read the story behind its development here.
Where do we download it from?
Please use the following link - Trinity
How do we run it?
Afterdownload, type make in the trinity base directory to compile Inchworm and Chrysalis. Butterfly, being written in Java, does not need any compilation. Then, run ./run_Trinity.sh on your reads with appropriate options. Please do not forget to add -run_butterfly option.
Is there a guide with step-by-step instructions?
Yes, it is right here.
Memory requirement
Trinity needs about 20-24Gb RAM for a library with 7.5M (100nt x2) paired end reads (please check our forum for an example).
Running Time
Trinity takes about 24 hours to assemble transcriptome from a sequence library with 7.5M (100nt x2) paired end reads (please check our forum for an example).
We have been writing on various topics for a while, but are not sure whether you find them useful. If you do not like to use public comment section, please feel free to email us at samanta at homolog.us.
