Viewing Insertion/Deletions

[Mindy]

1. The version of IGV: 2.2
   
2. Changed the View>Preferences>Mutations>color code mutations (to show Indels)
   
3. :
4. I thought that the coverage bars would change color to show that there was an indel/splice sight/ect but I still just see the single AA changes in a sample I know has a large and some small deletions.
5. Essentially what I would like to do is see where the INDELs are located vs SNPs just by looking at the colors. Thanks!!

Mindy

[Mindy]

Also, is there a way to set the threshold for insertion/deletion detection to 5% (that is what I have the threshold set at for SNPs).

[Jim Robinson]

Hi,

As you've noticed indels are not color-coded in the coverage plot,  so there is no reason to set a threshold.   We will consider this feature for a future release.

-- Jim

[Mindy]

I have a follow-up question: What does the DEL & INS number represent? It does not seem to be a count as the Total Count is less than that of the DEL:#. I am using IonTorrent data if that helps.

Total count: 307 A : 32 (10%, 28+, 4- ) C : 273 (89%, 156+, 117- ) G : 1 (0%, 1+, 0- ) T : 1 (0%, 0+, 1- ) N : 0 DEL: 4432 INS: 1 C>A 3106

Thanks! Mindy

[James Robinson]

The total count is the total number or bases aligned at that position.  Deletions and insertions do not represent bases at the position, rather the lack of bases or bases inserted to either side.   I see the confusion, perhaps we should draw a separator line under the last base count (the "N").

--
 
---
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/aLC7e7NI90g/unsubscribe?hl=en.
To unsubscribe from this group and all its topics, send an email to igv-help+u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Cheryl]

Hello,

I had a similar question to Mindy's.

Using her example, there are 4.432 deletions. Does this mean that in the alignment at that position there were 4,432 instances where, in order to get the best alignment, a gap was inserted at that position? & since deletions are represented by black bars, how do you get the total mumber of deletions at a given position?

Thanks!

[James Robinson]

I'm not sure I can address the alignment question,  however WRT the number of deletions at a given position you just draw a vertical line through that position and count how many black bars cross the line.   Not literally of course, but conceptually.

--

---
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/7c0cd8c7-d5fb-46c9-bb37-e559258fc113%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[bha...@pieriandx.com]

Hello,

IGV version: 2.3.72

I have similar question for insertion count. For a given position, the insertion count which is shown , does that show no. of reads which has a insertion just after the clicked position ?

For example,

I am looking at the variant from vcf file : chr5  VCF_Position: 112175969 Ref: G Alt: GA.

Now, If I click on the position 112175969, I am getting following counts:

Total count: 1017
A : 2 (0%, 0+, 2- )
C : 0
G : 1015 (100%, 543+, 472- )
T : 0
N : 0
---------------
DEL: 1

INS: 17  

But when I reviewed all the 17 reads which are showing the insertion, I found that out of 17 reads, 5 reads were having insertion of 'A' between the position 112175968 - 112175969. As per my understanding, the insertion count at the position 112175969, should give insertion count between the position 112175969-112175970. 

Please clarify , is it correct or not. If my understanding is correct, then is this a bug in IGV when it shows insertion count at a given position ?

Thank you for your time,

Waiting for your reply,

Bhakti

[James Robinson]

Hi, your understanding is correct,  the insertion is counted for the position preceding ("to the left") of the insertion.    I could not say if this is an igv bug or not without seeing the file.  Could you extract the portion of the file around the insertion?  You can do this from IGV by right-clicking and selecting export alignments.  I can send you a dropbox link to upload them.

Jim

[Bhakti Limaye]

Thank you Dr. Jim

I have shared my BOX a/c link with you via email.

There you will find bam file where I am looking at the insertion : chr5  VCF_Position: 112175969 Ref: G Alt: GA.

Out of 17 reads having insertions, 2 reads are with insertions "G" and "AGG" which are not the insertion which is mentioned in the vcf record.

In rest of 15, following are the reads which show insertion between 112175968-112175969 instead of insertion between 112175969-112175970.

The 5 read names are :

ReadName: VX6QE:02062:02662 

ReadName: VX6QE:02469:02737 

ReadName: VX6QE:02841:03082 

ReadName: VX6QE:03202:00889 

ReadName: VX6QE:03741:02800 

Please let me know your suggestions,

Thank you,

[James Robinson]

Got it.  I will try to look at it today.

[James Robinson]

Hi,

Thanks for the sample data.   If you notice the coverage track says there are 1,107 reads at this location,  but only ~50 or so are showing.   The rest have been downsampled (removed) for display.   The coverage track gives you the true count for every base.   For deep coverage the alignment track downsamples reads to a representative sample, otherwise you would run out of memory.   You can turn downsampling off in the alignment preferences to see the result.

However there is indeed an IGV bug here, or at least an inconsistency with the documentation.  Insertions on either side of the base selected are included in the count, not just insertions preceding the base.   I've attached a screenshot of the result with downsampling turned off,  16 reads have an insertion just before or after the centered location,  the final read is scrolled off view.

The downsampling indicator (dark bar under the coverage tracks) indicates the start location of downsampled alignments.   For the next release we will investigate a representation showing the complete coverage of downsampled alignments, rather than just start location, to make this clearer.   

Jim

[Bhakti Limaye]

Hello Dr.Jim

Yes I was aware of the settings of downsampling. So when I was working with this sample, I unchecked the option of down-sampling from  IGV and then marked those reads with 17 insertions. 

Thank you for clarification o this "However there is indeed an IGV bug here, or at least an inconsistency with the documentation.  Insertions on either side of the base selected are included in the count, not just insertions preceding the base." .

This will help to verify the alternate_allele count from my script and those which are observed on IGV.

However, I was unable to find attachment of the snapshot which you were mentioning in the comment .. 

Thank you so much for the explanation and guidance.

regards,

Bhakti

On Thursday, December 15, 2016 at 1:54:11 AM UTC+5:30, Jim Robinson wrote:

Hi,

Thanks for the sample data.   If you notice the coverage track says there are 1,107 reads at this location,  but only ~50 or so are showing.   The rest have been downsampled (removed) for display.   The coverage track gives you the true count for every base.   For deep coverage the alignment track down samples reads to a representative sample, otherwise you would run out of memory.   You can turn down-sampling off in the alignment preferences to see the result.

[James Robinson]

Sorry the screenshot didn't come through, it just showed there are 10 insertions on one side and 7 on the other.    This problem has been corrected for the next release.

Jim

[Mike Greenwood]

This is going back to the "N" question - If the DEL number is how many are missing, what could I use as the total in order to get a VAF for the position?  Total count before/after?  Average over the a few positions in the area?

Mike

[Mike Greenwood]

Maybe posting helped me think about it - should I do DEL/(DEL + Total count)?

[James Robinson]

I'm not sure what you're asking.  Are you asking how to compute a variant allele fraction?   If that's the case then yes,  I think DEL / (DEL + Total count) would be correct.

--

---
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/aLC7e7NI90g/unsubscribe.
To unsubscribe from this group and all its topics, send an email to igv-help+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/19192ad5-e09f-4a7c-a286-1d076d940ce4%40googlegroups.com.

[Louise C]

I have a question related to Mindy's.

Can you add the "total count" with the number of DEL and INS at that position?

The reason I ask is because I would like to know if it is possible to estimate the percentage of a deletion? Eg:

Total Count: 13069

A : 51 (0%, 43+, 8- )

C : 13008 (100%, 6936+, 6072- )

G: 8 (0%, 4+, 4-)
T : 2 (0%, 0+, 2-)

N : 0

----------------

DEL: 7539
INS: 0

Would it be correct to say that i have a 37% deletion 7538/(7539+13069)=0.3658x100= 37% at that position?

Thanks!

Louise

[James Robinson]

The total number of reads at that position is 13,069.   Of those,  7,539 have a deletion.      The total is not 7539 + 13069,  that is a meaningless number.

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+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/dc3becc2-1ee4-4172-9067-e8d083447e55%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Louise C]

Thank you. But I still wonder the following:

Is it possible to estimate a percentage of how much deletion you have at that position eg 7539/13069=58%

To unsubscribe from this group and stop receiving emails from it, send an email to igv-help+u...@googlegroups.com.

[Julius Fredens]

Hello and thanks for this useful group!

IGV_2.3.88

I notice that when I zoom out, insertions (purple I) disappear from the visual map. Is there a setting that allows insertions to be visible regardless of the zoom?

Cheers, Julius

[Helga Thorvaldsdottir]

Hi Julius,

Try the new IGV 2.4 beta and see if it helps. By default, the purple insertions are now visible when zoomed out to a much larger window than before. And if you find that they are too visible, you can use the Hide Indels option in the Alignment tab of the Preferences window.

Helga

[Julius Fredens]

Yep, that solves my problem. Many thanks, Helga

Julius