How to identify a gene targeted by external sequence into NGS data?

[aner]

Dear all
I am working on a Trascriptome project (Illumina GAIIx, paired ends).
I have 2 samples of Arabidopsis Thaliana. Within their sequences, an external sequence of a different genome has been inserted (transfection should be the word): the reason behind this choice is that we would able to identify the gene (prior unknown) that this sequence links, so that we can easily look at this gene expression.
My problem is that I need to identify which gene the sequence links.
Is there any software that achieves my goal?

I did classical RNA-seq steps:
- QC
- Alignment (Tophat - Bowtie)
- DEG (Cufflinks/diff)

I cannot search the external sequence into the aligned samples since the external sequence does not align with the samples' reference genome.

Thank you!!
Best
Andrea

[chadn737]

I'm not entirely clear on what the insertion is. Is it a T-DNA from Agrobacterium? That seems the most likely candidate since you are working with Arabidopsis.

The other problem is, do you know whether or not the insert is actually in a transcribed region?

If the sequence of insert is known, say if it is a T-DNA, then there are methods like TAIL-PCR that can be used to identify the flanking sequences of the insert.

[krobison]

I would probably just create a database from your possible insertion material & search it with your favorite short read search tool (e.g. BWA, Bowtie). That will identify any reads carrying the inserted DNA. If you can find a read pair (or do you not have them) with one read in the insertion and one in an Arabidopsis message, then you are home.

There are some tools out there specifically to look for these, but I forget the names.

[aner]

Thank you all for your messages!

Is it a T-DNA from Agrobacterium?

The other problem is, do you know whether or not the insert is actually in a transcribed region?

to chadn737: the insertion is GUS and the method is the gene trap. This time we tried without labeling a specific gene but letting GUS attach to all available genes.

There are some tools out there specifically to look for these, but I forget the names.

(...) just create a database from your possible insertion material & search it (...)

to krobison: your suggestion is really nice; if you figure out those tools, let me know ;-) Do you know how to create a database of reads? How can I call BWA/Bowtie to search for this insertion? Is it a Blast?

[krobison]

By database of possible insertion material I simply mean a FASTA format file with the possible insertions. The documentation for BWA, Bowtie & BLAST will instruct you how to index that file for searching (alas, each requires its own indexing).

BTW, a paper on finding such large insertions is

Identifying insertion mutations by whole-genome sequencing

[aner]

By database of possible insertion material I simply mean a FASTA format file with the possible insertions. The documentation for BWA, Bowtie & BLAST will instruct you how to index that file for searching (alas, each requires its own indexing).
BTW, a paper on finding such large insertions is
Identifying insertion mutations by whole-genome sequencing

Thank you very much. I will test this method and I will let you know results.

[bioinfosm]

as keith suggested, its straight forward to make a bwa formatted reference sequence (for the external sequence) and see what reads map to it... I hope you did paired reads, and the pair of the mapping read maps to the arabidopsis genome, so you can ascertain where the sequence links on the genome....

[aner]

I hope you did paired reads

Thank you bioinfosm. Yes, I have paired reads; I will see soon if the pairs maps to Arabidopsis genome. ...I hope to be lucky...!

[dcfargo]

You might try building a Bowtie index that includes the external sequence as an additional 'chromosome' and then running TopHat Fusion.

http://tophat-fusion.sourceforge.net/data/TopHatFusion-poster.pdf

[aner]

You might try building a Bowtie index that includes the external sequence as an additional 'chromosome' and then running TopHat Fusion.

Thank you dcfargo. I read the poster and I found the relative article (so far draft). I will try also this approach!