Closed marcosbarbeitos closed 1 year ago
Hi there,
I just tested MitoZ 3.6 with the test data (https://github.com/linzhi2013/MitoZ/wiki/Installation#9-running-the-test-dataset) on my Linux server, and it worked.
Probably it's the problem of Megahit. You may need to install Megahit by yourself and do a test run to check if your installation works or not, e.g., using the following command:
/path/to/your/own/megahit --out-dir ./megahit_out --num-cpu-threads 8 --k-list 43,71,99 --memory 500000000000.0 -1 /mnt/Data/local/mitoz3.6/test/clean_data/ttt.clean_R1.fq.gz -2 /mnt/Data/local/mitoz3.6/test/clean_data/ttt.clean_R2.fq.gz
If that's the case, replace the megahit-related binaries in /mnt/Data/local/mitoz3.6/bin/
with your newly installed ones.
Hello, it was indeed an issue with megahit, but it is not solvable by installation from source. My machine does not support POPCNT or BMI2, which seemingly leads megahit to use the following command line:
megahit_core_popcnt assemble -s /mnt/Data/local/mitoz3.6/megahit_out/tmp/k43/43 -o /mnt/Data/local/mitoz3.6/megahit_out/intermediate_contigs/k43 -t 8 --min_standalone 296 --prune_level 2 --merge_len 20 --merge_similar 0.95 --cleaning_rounds 5 --disconnect_ratio 0.1 --low_local_ratio 0.2 --cleaning_rounds 5 --min_depth 2 --bubble_level 2 --max_tip_len -1 --careful_bubble
This, however, does not work either.
After trying MitoFlex, I found out that by using megahit_core_no_hw_accel
instead of megahit_core_popcnt
I can get megahit to do its part.
I wonder if it would be possible to resume the assembly from this point by using a MitoZ option other that all
.
I imagine I'd also have to run megahit for k=71 and k = 99.
Thanks in advance!
If you want to continue the all
command after your own try of the megahit
command, you can add --resume_assembly
option.
if resume:
command = "megahit --continue --out-dir ./megahit_out "
if tmp_dir:
command += " --tmp-dir " + tmp_dir
runcmd(command, logger=logger)
this is because megahit supports this (--continue
).
If you don't want to do that, you can find the mt_assembly/megahit/megahit_out/final.contigs.fa
file, and search for mitogenome from it.
It has something like this:
>k99_363402 flag=1 multi=2.0000 len=304
If you want to use the findmitoscaf
command, it's better for you to reformat the mt_assembly/megahit/megahit_out/final.contigs.fa
file to this:
>k99_363402 abun=2.0000
Hello, I am testing Mitoz on an older Dell R900 server with 64GB of RAM and 16 cores running Scientific Linux ( = RHEL 7). After activating the conda environment, I am using the following command line:
However, I keep getting the following output, ending with an error message:
And this is MEGAHIT's log:
Any help will be greatly appreciated!
Cheers!