Open moldach opened 3 years ago
@moldach we haven't run celegans ourselves before. It looks like this might have something to do with chromosome names. Could you add the full list the chromosome names as they appear in your reference genome, mavis annotations file, and sample?
Hi @creisle
The names are roman numerals instead of numbers
I
II
III
IV
V
X
MtDNA
If it would be helpful I could share published sample for debugging
It took a couple months to figure out downloading the .json
annotations a while back (https://github.com/bcgsc/mavis/issues/205) but they look like they completed to me and I can find instances of each of these chromosome in it (with vim
).
I was working on the config parts of the code and thought of something we hadn't tried here. There's a parameter in clustering to configure what chromosomes to limit to but you can turn it off. Could you check the cluster logs for limit_to_chr
if that is set to the default human chromosomes that may be causing the issue.
Hi @creisle sorry about the delay. I've taken a look in the configuration logs for limit_to_chr
and I can confirm that I see human chromosome names here.
There's a parameter in clustering to configure what chromosomes to limit to but you can turn it off.
How do I do this? Is it a parameter I have to supply to mavis config
?
No worries, it's a busy time of year for all of us :)
You can either remove it post-config creation or you can set a new default by setting the environment variable pre-generating the config. Setting it to None will default to all chromosomes.
https://mavis.readthedocs.io/en/latest/configuration/settings/#limit_to_chr
export MAVIS_LIMIT_TO_CHR=None
You can see what all the defaults will be/are set to by generating a config file with the --add_defaults
flag
mavis config -w default_config.cfg --add_defaults
cat default_config.cfg
When you do this with the above environment variable set you should see this in the config it generates
limit_to_chr = None
Despite having a number of events (DEL, BND, DUP, INS) spread out across all chromosome in C. elegans (I, II, III, IV, V, X, MtDNA) I only see illustrations produced:
export MAVIS_DRAW_FUSIONS_ONLY=False
do I see ANY figures at all.I'm seeing this for a number of different strains (
BAM
//VCF' combos) I've run
mavis` on.To Reproduce Steps to reproduce the behavior:
My directory looks like:
The
output_dir
only contains the following images from X chromosomeConfiguration
Versions (please complete the following information):
CentOS Linux release 7.7.1908 (Core)
Python 3.8.2
mavis version 2.2.7
BLAT v. 35
Additional context
Furthermore, it looks like the majority of these figures are missing annotation information (since it says
_NA_NA.svg
)