Does the full length sequencing using Iso-Seq get the real full length ?

[huan]

I have used pacbio SMRT Iso-seq module to analyze my cDNA sequencing data, including Getting full length reads, Isoform level clustering (ICE and Quiver), Removing redundant transcripts. With the MatchAnno software, we do the isoforms annotiation analysis. But we find that, the isoform annoated to the known transcript is not full covered. Most of the isoforms overlap sequences lower than 50% because of the loss of 3' or 5' sequences. For example, we have an isoform: PB.338.1 with a length of 1979, the annoted transcript length is 3861, but the overlaped sequence can only be 808. I have confirmed that the 3`sequence has lost.
So, what's the matter with my sequencing data? Is the cDNA template is degraded? Or this is just a common problem with SMRT sequencing? How can I do with the anlysis of my data?

[bowhan]

The lost of 5' end sequence could be due to the degradation of the RNA sample, since the current protocol does not enforce an intact 5' end. There are methods to enrich intact RNA with 5' cap, for example, treating RNA with terminator 5' to 3' exonuclease to destroy degraded RNA with 5' phosphate.

On the contrary, the 3' end should be accurate because of the polyA selection (3' degraded RNAs without polyA should not be captured and cloned). For those missing 3' end, does the downstream sequence have sequences enriched in A (faked polyA)?

Also, did you do any size selection? If not, shorter RNAs are preferentially loaded and sequenced.

[huan]

Thanks a lot @bowhan
Yes. I have do the size selection. But we still can't fully sequence the entire lengths of long transcript. Trunction of sequence could occur form RNA defradation,mechanical shearing of the sample, incomplete PCR amplification, or loading bias of the sequencer.
So, IS there any software can be uesed to complete the sequence with illumina short reads by extending the ending of the sequences?
Thanks a lot again.

[nucacidhunter]

This product may enable full length transcript sequencing:https://www.lexogen.com/teloprime-fu...amplification/

[bowhan]

Thanks a lot @bowhan
Yes. I have do the size selection. But we still can't fully sequence the entire lengths of long transcript. Trunction of sequence could occur form RNA defradation,mechanical shearing of the sample, incomplete PCR amplification, or loading bias of the sequencer.
So, IS there any software can be uesed to complete the sequence with illumina short reads by extending the ending of the sequences?
Thanks a lot again.

I am not aware of any PacBio/illumina hybrid transcriptome assembler.

A CAGE library would be the most accurate method to determine the 5' ends/TSS. I would recommend GRIT (http://www.nature.com/nbt/journal/v3.../nbt.2850.html) for the analysis given that you already have RNA-seq.

But since you don't have a CAGE library yet, you might just as well build an Iso-Seq library with the key steps in the CAGE protocol to enhance 5' integrity (Terminator -> CIP -> TAP or its replacement -> 5' linker ligation -> RT with Iso-Seq RT primer -> PCR).
This paper might be helpful if you want to pursue this direction.

Gu, W. et al. CapSeq and CIP-TAP identify Pol II start sites and reveal capped small RNAs as C. elegans piRNA precursors. Cell 151, 1488-1500 (2012).

[huan]

Thanks a lot for your reply@bowhan @nucacidhunter.
That's quite a great idea to ensure the integrity of 5' end by library building.
But what I want is to deal with it by analysis method(software), such as extending the end of the transcript by assemble using short reads considering the alternative splicing condition. I am not sure is there any software or workflow could achieve this.
Any suggestion will be appreciated!