how many replicates are required for discovery of novel isoform using RNA-seq and cufflinks RABT assembly

Question

[rajeshkumar_vinod]

Hello everyone,

can I use RNA-seq data from 3 tissues each having 3,3,2 biological replicates and NO technical replicates to identify novel transcripts using cufflinks RABT based assembly?

if not then how many replicates do i need? Thank You.

[Carlo Yague]

I think that sequencing depth is more critical than replicates for novel transcripts discovery so only 3 or 2 replicates might not be the biggest issue for you.

[rajeshkumar_vinod]

Thanks for your reply.

yes it is paired end data and other things are also ok.

i was just thinking how can i say whether it is a novel transcript or not? If cufflinks reports this transcript in majority of biological replicates(eg. in 2 out of 3 or 3 out of 3 biological replicates it is present) only then?

[WouterDeCoster]

I agree with Carlo, depth will be more important than replicates. This paper might also be of interest to you: http://biorxiv.org/content/early/2016/01/29/038224

Answer 1

[EagleEye]

The isoform analysis itself is bit tricky. If you are going to discover some novel isoform, I would not recommend to just use cufflinks (though good for finding novel genes). You have to try using some specific method for isoform discovery, some useful posts

A: What is best tool to call diferentially expressed isforms ?

For sequencing I would recommend paired-end strand specific sequencing with more depth (This should compensate the need for more replicates).

[rajeshkumar_vinod]

Thanks for your reply.

yes it is paired end data and other things are also ok.

i was just thinking how can i say whether it is a novel transcript or not? If cufflinks reports this transcript in majority of biological replicates(eg. in 2 out of 3 or 3 out of 3 biological replicates it is present) only then?

and what is this long RNA-seq(>200 nt long)?

[EagleEye]

It has been discussed already and this post might be useful for you.

Rna-Seq: Novel Transcripts Found. What Next?

For long RNA-seq refer,

https://www.encodeproject.org/rna-seq/long-rnas/