Which transcriptome assembler is recommended for short-read cancer transcriptome data?

Question

[Benedek Dankó]

Hello!

I would like to do a de-novo transcriptome assembly on human control/cancer, short-read RNA-seq datasets. I would like to identify possible unannotated (cancer-specific) transcripts. Which assembler software would you recommend to use in this case?

Our server has 24 cores, 94 Gb RAM, so any softwares like Trinity would not be possible.

[JC]

Trinity is the best, it can reduce memory if you: - run samples separately - use normalization

Alternatively, rnaSPAdes can be used

Also, you can use some strategies to detect novel genes and fused genes using genome alignments.

[GenoMax]

any softwares like Trinity would not be possible

Trinity needs a GB of RAM per million reads. How many million reads do you have?

[Benedek Dankó]

Around 30 million (paired-end) per sample but there are samples with more (eg. 60-70 million).

[GenoMax]

You should be able to get them to run with hardware you have.