search

schneebergerlab / plotsr

Tool to plot synteny and structural rearrangements between genomes
MIT License
288 stars 28 forks source link

Can I use plotsr to visualize multiple human genomes with known SVs? #19

Closed calhoujd closed 2 years ago

calhoujd commented 2 years ago

Hi,

We have some cell lines where we made SVs using CRISPR editing. We would like to use plotsr to visualize these SVs, but I am struggling a bit to figure out how to do it. Unlike the example, I have bam files for 3 samples, but only a single reference .fa (the standard hg38.fa). I have plotsr installed and can get the example image, but not sure how to proceed with our data. Do I need to make genome assemblies from each sample (Illumina 150 bp WGS 30X) in order to run SiRY? It appears our HPC has SOAPdenovo installed, but it isn't clear to me that this outputs a .fa file...sorry, I am used to running packages that use Fastq or bam, definitely not used to using genome assemblies outside of the standard hg38 reference. Thanks in advance for any help!

Take care,

Jeff

PS - We have confirmed the SVs by both PCR and WGS with SV calling using Manta and lumpy.

bmansfeld commented 2 years ago

Hey Jeff I'm not part of the plotsr group (love the software btw) but I've had experience with validating SVs. I think you might want a different package to plot your SVs. I've used samplot (Belyeu et al., 2021) to plot the read coverage and identification of discordant mapping short reads. This will help you avoid assembling your new genomes (i doubt your 30X illumina would be good enough for that) Hope that helps

calhoujd commented 2 years ago

Hi Ben, Thanks for the suggestion to try samplot, I'll give it a try! I've been trying (so far unsuccessfully) to get either svviz or svviz2 up and running, lets hope installation of samplot is a bit more straight-forward. Thanks again! Take care, Jeff

mnshgl0110 commented 2 years ago

Hi Jeff, I would agree with Ben that in your setting, other packages might be more useful. Samplot is indeed cool and works very well. Also, sometimes, simple IGV screentshots are the best option.

I have a feeling that you are working with smaller SVs (10s to 100s bp long). Plotsr focus more at the chromosome level variations, so these SVs might not be visualised by it.

calhoujd commented 2 years ago

Hi Manish,

Our SVs are actually quite large (10+ kb) and probably would look quite nice with the plotsr visualization, but trying to do de novo assembly from our 30X Illumina genomes seems to be more effort than it is worth. Even if we can't use it for this, I did want to let you know that plotsr is a very nice package, congrats on it!

Take care,

Jeff