Finding Alternative splicing events from 454 transcriptome

[ketan_bnf]

Hi all.

I have transcriptome data from 454 sequencer. I have assembled it using de novo gsAssembler and formed a contigs (i am not using isotigs). I want to find alternative splicing events from these contigs, so far i am doing BLAT search for each contig as reference genome is known for my organism. Analyzing BLAT output in genome browser of UCSC comparing my contigs alignments with known mRNA, EST sequences whether splicing is different form reported mRNA splicing.

Is it the right approach?

the problem is i have to go one by one contig, it is taking to much time. Is there any other approach to speed up finding alternative splicing.

Thanks

[natstreet]

Can you explain why you are not running a transcript assembly and using isotigs as this would give you the information that you want.

[ketan_bnf]

thanks for reply natstreet

I have done a transcript assembly for transcriptome and formed isotigs. I am not much clear about isotig concept and also when analyzing isotig files, i can see just 4 bp as exons how can this be?. There are many such results in isotigs files. I had read about it like it is a bug in software, i am not sure. I am trying to understand isotigs, but till than i am using BLAT for alternative splicing.

[natstreet]

As a good starting point I can recommend reading through the posts at this blog. I think they will answer most of your questions but let us know if not.

[ketan_bnf]

According to this post isogroups contains isotigs which are formed from same contigs and also having variants

If i am not wrong these isotigs can be considered as variants of each other, How can i say which is best or important as splice variant?
from this, it says isotig02299 and isotig02300 are important as splice variants.

How can i relate this information to gene, do i have to do a blast/blat of isotigs to find which gene for they are coding?, how i should annotate this results?

[flxlex]

Some of the isotigs can be very similar to each other, with maybe a few bases differences. These are most likely caused by small variations between the transcripts from either the individual you are using (due to heterozygocity), or the mix of individuals. You could actually try to align all isotigs from one isogroup to check for this. Only the ones with a substantial length or sequence difference are candidate alternative splice variants.

If you have a reference genome you could align the isotigs to it to annotate them. Otherwise, blast/blat is not a bad starting point.

flxlex (author of the mentioned blog posts)