Non Uniform Gene Coverage And Cufflinks Predictions : Artifacts ?

Question

[Hugo D]

Dear all We are studying a 1500bp gene. Cufflinks on RNA-seq from 2 different biological groups predicts 3 splice forms: one full-length one and two shortened :

Maq mapping of reads confirms higher expression at the 5' and 3' ends. This supports Cufflink's predictions. first biological group : second biological group :

However, no reads indicate polyadenylation of the the 3' end of first short splice form. Thus we wonder if the 2 short splice forms are artifacts. Any ideas what may be going on?

hugo

[Michael]

Do you see polyA in other cases, known isoforms of other genes, alaways? Just wondering because some analysis steps might filter them. Further, does your RNA-seq protocol rely on polyA (oligo-dT primer) to fish mRNA, or random primers? If your protocol relies on polyA, how can this coverage be explained?

[Aaron Statham]

What scale are those graphs on? Also, it would be useful to see the individual read alignments

[Hugo D]

Hi

-> Each of these graphs is about 1000 reads of 75bp mapped on the full length transcript.

Read alignments on 200 bp regions (middle region vs end region) for second biological group :

-> You are right, we don't have polyA in other cases ! Poly(A) RNA is directly purified. Then after chemical fragmentation random primers are used to synthesized cDNA.

hugo

[Biomed]

Please edit your question and delete this "answer" instead of posting the additional information here as an "answer".

Answer 1

[Michael]

In conclusion, absence of polyA is no indication of anything, because your protocol wouldn't show it.

[Yannick Wurm]

okay - but is the weird coverage pattern due to the existance of 3 different spliceforms? or something weird?