search

zhenchengfang / PPR-Meta

A tool for identifying phages and plasmids from metagenomic fragments using deep learning
GNU Lesser General Public License v3.0
39 stars 9 forks source link

VM killed the process #6

Open Arzhura opened 4 years ago

Arzhura commented 4 years ago

Hello, I'm currently try to use the vitrual machine, everything is fine with the installation. I just have a problem, I try to launch the process on my metagenomes (size range between 136 159Ko and 725 892Ko), and everytime the process is killed. Is that due to the size of the metagenomes ? Thank you in advance.

zhenchengfang commented 4 years ago

Hi, will this problem happen when use the "example.fna" file? How many RAM have you allocated for the VM? Besides, the size of the host hard disk may also effect the VM (when the data size in VM exceeded the host hard disk).

Arzhura commented 4 years ago

Hi, no the problem does not occur when I try it with "example.fna". I allocated 2.5GB for the RAM, I can up it but for now I only have 6GB of RAM on my laptop. I just want to know if the tool has a limit size for file input. Then I will cut my files at this size. Thanks

zhenchengfang commented 4 years ago

Hi, the PPR-Meta tool itself dose not have a limit size for the input file, but the RAM that you allocate will affect the VM. 2.5GB is too little because the startup of MCR (one dependent software of PPR-Meta) already occupy 1~2GB of the RAM. I think that the RAM about 3GB larger than your input file maybe appropriate. Please contact me if there is another question!

Arzhura commented 4 years ago

Hi, I raise the memory to 4GB it is the more I can do with my laptop actually and I put the entire fasta file, not by block like I already done (it works), and it give me this output: image

So what is BLOCKREAD ? Thank you

zhenchengfang commented 4 years ago

Sorry for the late reply. "BLOCKREAD" is an option of "fastaread" function in MATLAB, which allow read the file in block. This error means that 4GB is still too little for your file. You can cut your file into smaller one. I tested the program with a 90.8 MB file and found that it could run in 4GB VM. Also, we found that run the software over a very large file would slow down the program. We will rewrite the code using "BLOCKREAD" option to reduce the requirement of the program and speed up the program in a few days. Thank you for your feed back.

Arzhura commented 4 years ago

Ok thank you, and I'm looking forward for the new version. I use your tool inside the virtual box but if it is possible to you to put the release in "release" in git I would try to reinstall on the cluster, I consider your tool as a submodule in my git, so the update is automatic if you put into "release" I think.

zhenchengfang commented 4 years ago

The new version (v1.1) can be downloaded from https://github.com/zhenchengfang/PPR-Meta/releases. You can replace the old folder with the new one in the VM. You can use the '-b' option to specify the number of sequences in each block, such as: ./PPR_Meta InputFile.fna result.csv -b 1000. We will update the "Manual.pdf" file as soon as possible. Please contact me if there is another question!