Using a MTAse Ko mutant instead of WGA
kafker opened this issue · comments
Dear devs,
I understand that nanodisco requires a WGA sample. However, I was wondering if a KO mutant for the methylase of interest (we already know the methylation motif) could be a "good" substitute for the WGA sample.
In theory, differences between the current of WT vs the current of the KO mutant strain should be observed only at the level of the methylation sites.
Perhaps I am missing something so I would like to hear your opinion.
Thank you!
Hello @kafker,
Yes, this is a perfectly good solution for identifying exact matches between methylation motifs and their Mtases. By running nanodisco
on this type of sample pair (i.e. native vs KO) you should mostly detect current differences at the KO motifs. Although be mindful that you might still see remaining of the other methylated motifs as methylation level could slightly variate between both samples. This effect should be minimal.
I'll be interested in hearing back from you if you pursue this type of experiment as we made only few of this type of experiment.
Best,
Alan
Hello @kafker,
Yes, this is a perfectly good solution for identifying exact matches between methylation motifs and their Mtases. By running
nanodisco
on this type of sample pair (i.e. native vs KO) you should mostly detect current differences at the KO motifs. Although be mindful that you might still see remaining of the other methylated motifs as methylation level could slightly variate between both samples. This effect should be minimal.I'll be interested in hearing back from you if you pursue this type of experiment as we made only few of this type of experiment.
Best,
Alan
Hi Alan,
I will keep you updated
Best
Hi Alan,
Just a quick update. I am struggling quite a bit with nanodisco difference
because I can't find a good way to speed up the process. Here is the batch script
#!/bin/bash
#SBATCH --job-name="nanodisco_diff_gzip_100k_8"
#SBATCH -A Account Name
#SBATCH --partition p_name
#SBATCH -t 96:00:00
#SBATCH --mem=300G
#SBATCH --qos=p_qos_name
#SBATCH --ntasks=6
#SBATCH --cpus-per-task=7
#SBATCH --nodes=1
#SBATCH --output=nanodisco_diff_gzip_100k_8_log.txt
#SBATCH --error=nanodisco_diff_gzip_100k_8_err.txt
export TMPDIR=$SCRATCH/ONT/singularity/nanodisco/tmp_8
singularity exec --bind /path/to/ONT nanodisco.sif nanodisco difference \
-nj 6 -nc 1835 -p 7 -i /path/to/ONT/singularity/nanodisco/preprocessed_all_gzip \
-o /path/to/ONT/singularity/nanodisco/difference_gzip_100k_8 -w KO_NAT -n WT_NAT \
-r /path/to/ONT/ref.fasta
The computing node that I am currently using has 2 x CPU with 24 cores each, 2.4 GHz, and 384GB RAM.
In trying to run nanodisco difference on a fill genome (1835 chunks) and in two days I got only 460 chunks.
Probably I will need to relaunch the analysis on a node without a maxwalltime
. Let see...
In the meanwhile, I used deepmod2 on the WT and KO mutant and, as expected, I am observing huge differences in the read methylation fraction at the level of Mtase methylation motifs. Since we do not have a WGA, I think It would make more sense to subset of the reference genome to specifically targeted the regions of interest.
I will keep you updated and thank you for your support!
I used nanodisco for WT and MTase mutant, it worked well.