IGV junction/sashimi plots for chimeric alignment

345 views
Skip to first unread message

Jinhyuk Bhin

unread,
Sep 21, 2020, 1:28:59 PM9/21/20
to igv-help
Hi,

I am using IGV to visualize RNA-seq driven rearrangment,

One of the interesting rearrangement we are interested is the one that the end of certain exon of a gene (- strand) is linked to intergenic region before that gene in opposite strand  (+ strand). Apparently, there are many reads supporting the reads but I cannot see any junction connection in the sashimi plot.

Does Junction plot only visualize the junction in the same strand?


James Robinson

unread,
Sep 21, 2020, 1:30:23 PM9/21/20
to igv-help
All junctions are shown, for both strands.

--

---
You received this message because you are subscribed to the Google Groups "igv-help" group.
To unsubscribe from this group and stop receiving emails from it, send an email to igv-help+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/c621a8d0-47c5-45f1-8484-0647b7cf29f5n%40googlegroups.com.

Jinhyuk Bhin

unread,
Sep 21, 2020, 1:33:17 PM9/21/20
to igv-...@googlegroups.com

Thanks for the quick reply.

 

 

If you see the plot, there is no junction connection, but there are junction reads in alignment panel.

How can I understand this?

--

---
You received this message because you are subscribed to a topic in the Google Groups "igv-help" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/igv-help/so1Z627TwtI/unsubscribe.
To unsubscribe from this group and all its topics, send an email to igv-help+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/CACOP%2BpuOh%2B_JE3Zs5mNyxfH0eHvo2tRW7jXB-fWF62AdsR8kEA%40mail.gmail.com.

James Robinson

unread,
Sep 21, 2020, 1:39:59 PM9/21/20
to igv-help
I don't understand this picture.  Is this a paired-end read view?   I do not believe those are junctions  (technically gaps marked "N" in the cigar string).    Junctions only occur within a single read, by definition, and a single read can have only 1 strand.

Jinhyuk Bhin

unread,
Sep 21, 2020, 1:48:22 PM9/21/20
to igv-...@googlegroups.com
If there is a flipped rearreangement between the intron of that gene and somewhere intergenic region, wouldn't it be possible to make the fused transcript from that exon (- strand) to pseudo exon (+) strand in the intergenic region?

2020년 9월 21일 (월) 19:40, James Robinson <jrob...@broadinstitute.org>님이 작성:
--

---
You received this message because you are subscribed to a topic in the Google Groups "igv-help" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/igv-help/so1Z627TwtI/unsubscribe.
To unsubscribe from this group and all its topics, send an email to igv-help+u...@googlegroups.com.

James Robinson

unread,
Sep 21, 2020, 1:52:30 PM9/21/20
to igv-help
Biologically yes, but a "junction" in the parlance of IGV (and other tools) is a gap in a single read.   The picture you show connects 2 different reads, it's not possible for a single read to align to 2 strands unless it is split, which in the BAM spec is represented by different reads.   So I think the picture you are showing must be paired-end reads, the line represents connections between the 2 pairs,  not a gap in a single read.   That's why I asked you if this is a paired-end view  ("view as pairs").



You received this message because you are subscribed to the Google Groups "igv-help" group.
To unsubscribe from this group and stop receiving emails from it, send an email to igv-help+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/CADFumqqYdoHhz37S4Pq%2BsJrduB5d8hETCvzAWi8mDUhoViLtxA%40mail.gmail.com.

Jinhyuk Bhin

unread,
Sep 21, 2020, 2:50:46 PM9/21/20
to igv-...@googlegroups.com

Yes, you are right. Those have gray line which indicates two pairs of reads.

 

The issue comes from STAR alignment that I used. Actually, these are chimeric sequences that R1 was derived from the chimeric reads spanning the exon (- strand) to the intergenic (+ strand) which were split into two alignment by STAR (one only for exonic region and the other is only for intergenic region) and R2 was derived from the normal continuous reads in the intergenic reads.

 

STAR split this R1 into two different alignments making one to be supplementary alignment (because of opposite strand mapping?).

 

Actually, what I want to do is making the sashimi plot that supports such chimeric alignments from the exon to intergenic region.

 

I now know that this is the issue of STAR, but is there any suggestion that I can make such sashimi plot supporting the chimeric alignments in the different strand?

 

Thanks

 

 

 

 

From: <igv-...@googlegroups.com> on behalf of James Robinson <jrob...@broadinstitute.org>
Reply-To: <igv-...@googlegroups.com>
Date: Monday, 21 September 2020 at 19:52
To: igv-help <igv-...@googlegroups.com>
Subject: Re: [igv-help] IGV junction/sashimi plots for chimeric alignment

 

Biologically yes, but a "junction" in the parlance of IGV (and other tools) is a gap in a single read.   The picture you show connects 2 different reads, it's not possible for a single read to align to 2 strands unless it is split, which in the BAM spec is represented by different reads.   So I think the picture you are showing must be paired-end reads, the line represents connections between the 2 pairs,  not a gap in a single read.   That's why I asked you if this is a paired-end view  ("view as pairs").

 

 

 

On Mon, Sep 21, 2020 at 10:48 AM Jinhyuk Bhin <bhin...@gmail.com> wrote:

If there is a flipped rearreangement between the intron of that gene and somewhere intergenic region, wouldn't it be possible to make the fused transcript from that exon (- strand) to pseudo exon (+) strand in the intergenic region?

 

2020 9 21 () 19:40, James Robinson <jrob...@broadinstitute.org>님이 작성:

I don't understand this picture.  Is this a paired-end read view?   I do not believe those are junctions  (technically gaps marked "N" in the cigar string).    Junctions only occur within a single read, by definition, and a single read can have only 1 strand.

 

Thanks for the quick reply.

 

cid:174b1bab22c4cff311

 

If you see the plot, there is no junction connection, but there are junction reads in alignment panel.

How can I understand this?

 

 

--

---
You received this message because you are subscribed to a topic in the Google Groups "igv-help" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/igv-help/so1Z627TwtI/unsubscribe.
To unsubscribe from this group and all its topics, send an email to igv-help+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/CACOP%2BpsfJTB76JyD7urMhu5xD_RsH6BGe6PuD9OFc7_4nkVEqQ%40mail.gmail.com.

--

---
You received this message because you are subscribed to the Google Groups "igv-help" group.
To unsubscribe from this group and stop receiving emails from it, send an email to igv-help+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/CADFumqqYdoHhz37S4Pq%2BsJrduB5d8hETCvzAWi8mDUhoViLtxA%40mail.gmail.com.

--

---
You received this message because you are subscribed to a topic in the Google Groups "igv-help" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/igv-help/so1Z627TwtI/unsubscribe.
To unsubscribe from this group and all its topics, send an email to igv-help+u...@googlegroups.com.

James Robinson

unread,
Sep 21, 2020, 3:03:22 PM9/21/20
to igv-help
Hi,

First do you mean sashimi or junction plots?  The figures you have show junctions.   You can't do this directly from the BAM file, but you  can create a junction track from a from "bed" file, where each bed record represents a junction, you need a track line with "graphType=junctions".

You can't create a Sashimi plot this way with IGV, there are other Sashimi packages out there,  do a web search, you might find one that can help.   

Jinhyuk Bhin

unread,
Sep 21, 2020, 3:07:52 PM9/21/20
to igv-...@googlegroups.com

What I meant was sashimi plot supporting chimeric reads. But if STAR first treat this read as two different alignment based on the different strand, it should be solved first how to deal with this.

 

I’ll do web search to solve this problem!

 

Thanks!!

From: <igv-...@googlegroups.com> on behalf of James Robinson <jrob...@broadinstitute.org>
Reply-To: <igv-...@googlegroups.com>
Date: Monday, 21 September 2020 at 21:03
To: igv-help <igv-...@googlegroups.com>
Subject: Re: [igv-help] IGV junction/sashimi plots for chimeric alignment

 

Hi,

--


---
You received this message because you are subscribed to a topic in the Google Groups "igv-help" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/igv-help/so1Z627TwtI/unsubscribe.
To unsubscribe from this group and all its topics, send an email to igv-help+u...@googlegroups.com.

Reply all
Reply to author
Forward
0 new messages