SEQanswers

SEQanswers (http://seqanswers.com/forums/index.php)
-   Bioinformatics (http://seqanswers.com/forums/forumdisplay.php?f=18)
-   -   RNAseq alignment -misaligned on the genome! (http://seqanswers.com/forums/showthread.php?t=75138)

sitapriyamoorthi 03-31-2017 09:13 AM

RNAseq alignment -misaligned on the genome!
 
Hello ,

My colleague did the mapping of publicly available human RNAseq data. However when I view this alignment on IGV the reads map next to the genes and not above it? Why would this be happening? Any suggestions?

GenoMax 03-31-2017 09:27 AM

Can you clarify what do you mean by that? Perhaps attach a screenshot to explain.

Bukowski 03-31-2017 12:09 PM

You're looking at the wrong genome build/annotation in IGV almost certainly.

r.rosati 03-31-2017 01:28 PM

I agree with Bukowski, if you're visualizing the data against hg19, try switching to GRCh38 - or vice-versa. Your collaborator aligned the data against a genome build that's different from the one you're using on IGV.

https://i.imgur.com/mw3V07T.png

wdecoster 03-31-2017 01:37 PM

Classic mistake, but you'll make this one only once ;)

sitapriyamoorthi 03-31-2017 02:38 PM

Nope correct genome build
 
Nope we are looking at it with the same genome build that he used to align same problem.

r.rosati 03-31-2017 05:23 PM

Curious... can you provide a screenshot?

sitapriyamoorthi 04-01-2017 11:41 AM

Here are the screen shots
 
2 Attachment(s)
So this is the screen shot for ACTB gene and as you can see the reads show up ~1kb downstrem

r.rosati 04-04-2017 01:25 PM

1 Attachment(s)
I have half an hypothesis. See if you agree.
By the (blurry) coordinates it seems like you're using hg19, is this correct? If so, then ACTB is within the window chr7:5,566,700-5,570,300. I'll use these numbers.

First thing, if your IGV is set to show mismatches in color, then you will agree with me that the build used for alignment and the build you're using on IGV don't match, because the nucleotides are all in color = mismatched. On the other side, if you've set IGV to show all nucleotides in color regardless of match, then disregard this comment.

You do see some reads about 1kb downstream, but you will agree with me that the exon structure totally doesn't match that of ACTB. You see a long exon, then three short ones, then one last long exon. This doesn't match ACTB.

However, 22kb downstream you have the FSCN1 gene, whose exon structure really seems to match the transcriptome reads. I'm not affirming those reads totally absolutely belong to FSCN1, but...

Attachment 4738

If this is the case, and you say that you see these reads 1kb downstream to ACTB, then maybe your reads are actually shifted 60kb upstream.
And if so, then you should see transcriptome reads matching the ACTB exon structure within about the region: chr7:5,500,000-5,518,000.

The hole in this hypothesis is: I don't know which genome build would have ACTB shifted 60kb upstream compared to hg19. It's not GRCh38 nor hg18. Could it be an earlier version of hg19?

In the end, I would talk to your collaborator and be 100% sure about the genome build he's used to align the data. Although the mapping doesn't match hg19 the data is likely not wrong, but it's best to realign it to GRCh38 or hg19.

r.rosati 04-13-2017 12:33 PM

...Any news?

sitapriyamoorthi 04-14-2017 06:03 AM

Dear All,
I thought I had replied to this thread but I hadn't. R.Rosati, thank you for you painstaking response. I asked my collaborator again to check the genome build he had used. Being a yeast person he did not realize how different genome annotations were version to version, infact he had used GRCh38 and not hg 19. Problem resolved! Thank you all for your prompt responses it helped me query the problem with my collaborator more confidently!


All times are GMT -8. The time now is 08:40 AM.

Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2019, vBulletin Solutions, Inc.