[BioC] Visualization of alignments with mismatch bases

[Julian Gehring]

Hi Tengfei,

Thanks a lot for taking this into account!  I'm sure many users would 
find it useful.

Do you have any code drafts on how to plot alignments and mismatches?  I 
know doing this will be hard, but at the moment I would be willing to 
give it a try.

Best wishes
Julian


On 02/28/2013 10:05 PM, Tengfei Yin wrote:
> Thanks Julian, for you feedback, I will keep your feature request in my
> mind, and will think about implementing something like different
> modes/types in geom_alignment or stat_mismatch function in ggbio.
>
> Tengfei
>
> On Thu, Feb 28, 2013 at 2:31 PM, Julian Gehring
> <julian.gehring at gmail.com <mailto:julian.gehring at gmail.com>> wrote:
>
>     Hi Tengfei,
>
>     What you describe covers what I tried to do so far (overlapping
>     tracks, really a hacky task).  Having these modes would be very
>     handy to have, especially in combination with the other capabilities
>     of 'ggbio'.  All the three modes would be good to have; I would
>     consider (1) the most useful one for me at the moment.  The more
>     detail to incorporate in the alignments (e.g. including SNV
>     information of a VCF file) would need additional relations between
>     BAM alignments and the VCF.
>
>     I am not aware of any R package with this functionality, but I see a
>     large potential for this.  Currently, I interact with IGV using
>     batch mode, but this far away from being a good and long-term solution.
>
>     Best wishes
>     Julian
>
>
>
>         You are right, currently ggbio only supports summary of mismatch
>         showing
>         as coverage plot and barchart(?stat_mismatch), but looks like
>         what you
>         want is detailed short reads alignments visualization with mismatch
>         bases showing right on the reads, . It's possible, but not easy
>         to do it
>         manually...you have to have two GRanges objects, one for
>         alignment one
>         for SNP, and plot them layer by layer, the tricky part is
>         assigning each
>         reads fixed stepping level, so snp can be plotted on the right
>         position.
>         I will NOT recommend you to do this, it's probably not worth
>         taking time
>         doing it. I need to implement this features in some easy way.
>
>         The tricky part is that there are different modes, 1. show reads
>         as gray
>         rectangle, and color mismatch as segment  2. show SNP as nucleotide
>         text, A/C/T/G..,  3. show sequence detail for each alignment. those
>         depends on zoomed level and even coverage, and I guess most time you
>         don't want to see bases for every reads...
>
>         Just curious for future ggbio development, are those modes want you
>         want?  are you just using bam files here? no VCF files involved
>         right?
>         Because you mentioned 'snp', I think what you mean is mismatch?
>
>         ps:  I cannot speak for other tools, and only thing I know, in SRAdb
>         package, looks like it could fire your data in IGV..
>
>         Thanks
>
>         Tengfei
>
>
>
>
> --
> Tengfei Yin
> MCDB PhD student
> 1620 Howe Hall, 2274,
> Iowa State University
> Ames, IA,50011-2274
>
>