Question: How To Distinguish Chimeric Transcript From Fusion Gene
3
gravatar for upendrakumar.devisetty
6.4 years ago by
United States
upendrakumar.devisetty350 wrote:

Can someone tell me how to distinguish chimeric transcript from fusion gene? I am very confused.

I have recently made a transcriptome assembly from Vevelt/Oases pipeline and blasted it against Ref genome and i found that some of the transcripts were blasted on two different chromosomes and i also found that some of the transcripts blasted on the same chromosome but at different locations. Now i am wondering if i need to keep these chimeric transcripts in my final assembly or throw them away?

Thanks Upendra

gene fusion • 4.2k views
ADD COMMENTlink modified 6.4 years ago by cdsouthan1.8k • written 6.4 years ago by upendrakumar.devisetty350

Can you clarify what you mean by a "fusion gene"? Searching for chimeras is fairly straightforward. There could be many different reasons why your reads are hitting different locations.

ADD REPLYlink modified 6.4 years ago • written 6.4 years ago by Josh Herr5.6k
1

By fusion gene i mean transcripts originating from two different parts of the genes either by translocation, deletion etc., This is biological. What i am worried is if i throw away the chimera transcripts i would be throwing away biologically interesting genes. On the other hand if i keep those chimera transcripts i am worried that it will be cause major artefacts in analysis following a transcriptome assembly, like detection of sequence or expression variation.

ADD REPLYlink written 6.4 years ago by upendrakumar.devisetty350

Probably splitting the problem in two and solving each one separately would bring the best results!

1) Use first Velvet/Oases and throw away those transcripts which map on two different chromosomes (most likely due to assembly errors there are many false positives fusion genes)

2) For finding fusion genes use specificaly design tools like for example FusionCatcher http://code.google.com/p/fusioncatcher/ (it has very good sensivity and specificity for finding fusion genes)

3) use the results from step 1 and step 2 together

ADD REPLYlink modified 6.4 years ago • written 6.4 years ago by Enx20

This is common to de novo assembly. For a large genome, most of chimeric contigs are caused by misassembly (including misassembly in the reference genome) instead of anything biological.

ADD REPLYlink modified 6.4 years ago • written 6.4 years ago by lh331k

I agree with you regarding reference genome. How about denovo assembly? How about transcriptome assembly? Keep or not keep?

ADD REPLYlink written 6.4 years ago by upendrakumar.devisetty350
2
gravatar for cdsouthan
6.4 years ago by
cdsouthan1.8k
cdsouthan1.8k wrote:

As Josh says, in terms of causes of chimeras as phenomena they could be technical (e.g. library preparation) and or biological reasons, particularly associated with rare "events" (see http://www.ncbi.nlm.nih.gov/pubmed/11840564). For example if one animal used for the cDNA library has a chromsomal abnormality a fusion gene is exactly what you might get and a library from a cancer cell line could give you loads of them. Nominally if you want a reference set of contiged transcripits its your choice what to reject but remember the misfits may be rare but biologically real.

ADD COMMENTlink modified 6.4 years ago • written 6.4 years ago by cdsouthan1.8k

Thanks. You think i should keep them ?

ADD REPLYlink written 6.4 years ago by upendrakumar.devisetty350

As ever its your call. If the experiment with your collaborators produced a good and unique cDNA library (what organism/tissue BTW?) with deep and high quality reads then its worth working hard at contiging the transcripts and submitting to TSA. You are in a double-bind as pointed out because chimeras by any cause will confound transcript contiging and of course any draft genomic assembly will have its own assembly artefacts (e.g. split genes) that confound the mapping of some transcripts anyway. Just keep the misfits transcript file for a rainy day... Alternatively you could experiment with coverage thresholds, ie a putative chimeric transcript supported by many reads is more likely to be "interesting"

ADD REPLYlink modified 6.4 years ago • written 6.4 years ago by cdsouthan1.8k

Sounds good cdsouthan. My denovo transcriptome assembly reads came from deep sequencing of a plant (Brassica rapa) from very many tissues (9 different in total). I could easily look to see the coverage for those putative chimeric transcripts and then i will decide if i want to keep them or not. Thanks anyway for your suggestion.

ADD REPLYlink written 6.4 years ago by upendrakumar.devisetty350
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1783 users visited in the last hour