Re: vcfanno : LUMPY and SV annotations

[Bogdan Tanasa]

Hi Brent,

talking about a vcf file from LUMPY, please would you advise if my interpretation of the results (below) is correct. Thank you !

the files that I am using are the following :

-- input : "conf.toml" and "vcf.LUMPY.example.vcf", the TRA are listed on 2 consecutive lines (for both breakpoints) : eg a chr11-chr22 translocation is :

chr11    128801587    12410_1    
chr22    29292248    12410_2

-- output : "vcf.LUMPY.example.ANNO.vcf" where each breakpoint is annotated on a separate line (i.e. a BND is indicated on 2 consecutive lines).

bogdan

 

On Mon, Feb 5, 2018 at 8:46 PM, Bogdan Tanasa <tan...@gmail.com> wrote:

Dear Brent, thank you for the very quick and informative reply.

hmmm, somehow it works (in some extent) with the BND data from LUMPY, when VCFANNO generates a label for each breakpoint,

although I will have to check with you that it is correct. i will email you in 10-15 minutes, and please let me know what you think ...

On Mon, Feb 5, 2018 at 8:38 PM, Brent Pedersen <bped...@gmail.com> wrote:

Hi Bogdan,
no, vcfanno won't annotate that correctly. First, <TRA> is not valid
VCF ALT as it's not in the spec. Also, vcfanno can not annotate across
chromosomes at this time.
Given the way that it works internally, it would be hard to annotate
interchromosomal events like this...
-Brent

On Mon, Feb 5, 2018 at 9:09 PM, Bogdan Tanasa <tan...@gmail.com> wrote:
> Dear Brent, thank you again for your very precious help.
>
> In principle, i've got vcfanno to run, and to annotate a VCF file with SV
> information ; there is a specific issue I am encountering though :
>
> the SV vcf file has been generated by DELLY 0.7.5, and an entry for TRA/BND
> looks in the following way (ie. the INFO field):
>
> chr22    29292248    TRA00070224    A    <TRA>    .    PASS
> PRECISE;SVTYPE=TRA;SVMETHOD=EMBL.DELLYv0.7.5;CHR2=chr11;END=128801587;INSLEN=0;HOMLEN=2;PE=20;MAPQ=60;CT=5to3;CIPOS=-2,2;CIEND=-2,2;SR=10;SRQ=1;
>
> The question would be : would VCFANNO recognize in the vcf file the fields
> CHR2 and END (above) as the other side of the translocations ?
> i could send you an example in the next email (and will post the question
> online too), as I can annotate only one end of the translocation ...
>
> thank you,
>
> -- bogdan
>
>
> On Sun, Feb 4, 2018 at 7:48 AM, Brent Pedersen <bped...@gmail.com> wrote:
>
>>
>> Hi Bogdan,
>> Glad it was helpful. I am happy to help with other questions, if you
>> can, it would be helpful if you ask them on the vcfanno mailing list
>> so that others can benefit from the information.
>>
>> -Brent
>>
>> On Fri, Feb 2, 2018 at 11:49 AM, Tanasa, Bogdan <Bogdan...@ucsf.edu>
>> wrote:
>> > Dear Brent, thank you for your help -- it is very very precious.
>> >
>> >
>> > I will use a vcf file from DELLY and a GFF or BED file with the gene
>> > annotations on hg38. I may need a bit of extra-help with setting up the
>> > conf.toml, and will let you know how it goes.
>> >
>> >
>> > with much appreciation,
>> >
>> >
>> > -- bogdan
>> >
>> > ________________________________
>> > From: Brent Pedersen <bped...@gmail.com>
>> > Sent: Friday, February 2, 2018 10:40:44 AM
>> >
>> > To: Tanasa, Bogdan
>> > Cc: tan...@gmail.com
>> > Subject: Re: vcfanno
>> >
>> > If the gene name and strand are in your BED file, you can tell vcfanno
>> > to report those, just as you normally would, e.g. something like:
>> >
>> > columns=[4, 5]
>> > names=["gene_name", "gene_strand"]
>> >
>> > On Fri, Feb 2, 2018 at 11:29 AM, Tanasa, Bogdan <Bogdan...@ucsf.edu>
>> > wrote:
>> >> Dear Brent,
>> >>
>> >>
>> >> thank you for your email and for all the information. Please may I also
>> >> ask
>> >> you :
>> >>
>> >> when using vcfanno to annotate the SV, for a breakpoint, could we have
>> >> both
>> >> the overlapping GENE information and the STRAND of the GENE (direction
>> >> of
>> >> transcription) ? If it is so, I could infer the FUSIONS / TRUNCATIONS
>> >> ...
>> >>
>> >>
>> >> Please if it is possible to share an example of the OUTPUT for CNV or
>> >> SV
>> >> annotations, it would be very very helpful. Thanks a lot !
>> >>
>> >>
>> >> thank you,
>> >>
>> >>
>> >> bogdan
>> >>
>> >>
>> >>
>> >> ________________________________
>> >> From: Brent Pedersen <bped...@gmail.com>
>> >> Sent: Friday, February 2, 2018 9:36 AM
>> >>
>> >> To: Tanasa, Bogdan
>> >> Cc: tan...@gmail.com
>> >> Subject: Re: vcfanno
>> >>
>> >> There are some docs here: https://github.com/brentp/vcfanno#-ends
>> >>
>> >> but no, it cannot annotate fusions and annotations.
>> >>
>> >> On Fri, Feb 2, 2018 at 9:54 AM, Tanasa, Bogdan <Bogdan...@ucsf.edu>
>> >> wrote:
>> >>> Dear Brent,
>> >>>
>> >>>
>> >>> thanks a lot for a very quick and helpful reply ! please may I also
>> >>> ask
>> >>> where could I find the latest documentation on the use of vcfanno on
>> >>> CNV
>> >>> and
>> >>> SV annotations. Could I only input a set of RefSeq genes and obtain
>> >>> also
>> >>> the
>> >>> information related to fusions and truncations ? many many thanks
>> >>> again,
>> >>>
>> >>>
>> >>> -- bogdan
>> >>>
>> >>> ________________________________
>> >>> From: Brent Pedersen <bped...@gmail.com>
>> >>> Sent: Friday, February 2, 2018 6:35:13 AM
>> >>> To: Tanasa, Bogdan
>> >>> Cc: tan...@gmail.com
>> >>> Subject: Re: vcfanno
>> >>>
>> >>> Hi Tanasa,
>> >>> yes, vcfanno works for annotating SVs and CNVs, it uses the END or
>> >>> SVLEN fields from the INFO if available and automatically switches
>> >>> to overlap testing (instead of requiring exact match to REF and ALT as
>> >>> is the default).
>> >>> There's also a special flag in vcfanno `-ends` that will annotate the
>> >>> left and right ends of the SV in addition to the full interval as we
>> >>> are
>> >>> often interested in where the breakpoint falls.
>> >>> hope that helps,
>> >>> -Brent
>> >>>
>> >>> On Fri, Feb 2, 2018 at 2:13 AM, Tanasa, Bogdan
>> >>> <Bogdan...@ucsf.edu>
>> >>> wrote:
>> >>>> Dear Brent,
>> >>>>
>> >>>>
>> >>>> greetings from UCSF, where we are working on a clinical genomics
>> >>>> project
>> >>>> that aims to do whole genome sequencing of  pediatric cancers.
>> >>>>
>> >>>>
>> >>>> When you have a minute, I would like to ask you please a quick
>> >>>> question
>> >>>> :
>> >>>> could we use Vcfanno for the annotations of SV and CNV ?
>> >>>>
>> >>>>
>> >>>> thank you,
>> >>>>
>> >>>>
>> >>>> -- bogdan
>> >>>
>> >>
>> >
>
>

[Brent Pedersen]

Hi Bogdan, I'm not sure what your question is.

In this case, if the POS and the END (from the INFO) are for the region of interest for a given line, then vcfanno will annotate it as expected.

-Brent