Install Juicer please

[NoeliaDiaz]

👤 Your Marbits Username

ndiaz

🖥️ App Name

Juicer

🔗 App URL

https://github.com/aidenlab/juicer

📦 App Version

Juicer 1.6

📄 Main Publication

Neva C. Durand, Muhammad S. Shamim, Ido Machol, Suhas S. P. Rao, Miriam H. Huntley, Eric S. Lander, and Erez Lieberman Aiden. "Juicer provides a one-click system for analyzing loop-resolution Hi-C experiments." Cell Systems 3(1), 2016.

📝 Additional Notes

[franlat]

Installed from source!

Should work with

module load juicer/1.6

Juicer can use GPU, but we do not have GPU nodes in Marbits. If it fails, we can see if we can install the CPU alternative.

[NoeliaDiaz]

Hi,
I tried to run Juicer several times now, and it was breaking. At first, I thought it was my slurm and changed it several times, but after carefully checking the logs, I think the problem is a module juicer/1.6 installation issue that prevents Juicer from completing any pipeline after the mapping.

Juicer is installed in: /home/apps/juicer/1.6/ , and it is incomplete. It lacks directories and scripts that are essential for Juicer to run merge, dedup, final, and postproc; the module points to routes that don't exist.

Example of errors:
awk: fatal: can't open source file
/home/apps/juicer/1.6/CPU/scripts/common/chimeric_blacklist.awk
(No such file or directory)

The same happens for:
stats_sub.awk
collisions.awk
collisions_dups.awk
collisions_dedup_rearrange_cols.awk
juicer_tools
juicer_postprocessing.sh
statistics.pl
check.sh

Because all these routes are at: /home/apps/juicer/1.6/CPU/scripts/common/ that does not exist in the current installation.

It should contain:
CPU/
juicer.sh
common/
scripts/
common/
SLURM/
scripts/
common/

But right now at Marbits we only have:
/home/apps/juicer/1.6/CPU/common/
/home/apps/juicer/1.6/SLURM/scripts/

The only solution that I think of is to reinstall it from github:
git clone https://github.com/aidenlab/juicer.git
bash scripts/install.sh

or since you have sudo powers...hahaha, maybe just copy the missing directories to:
CPU/scripts/common/
CPU/scripts/
SLURM/scripts/common/

[franlat]

I am trying to understand what is happening, because those scripts are in the installation, but not in the path that juicer 1.6 is complaining about.

So, this does not exist: /home/apps/juicer/1.6/CPU/scripts/common/
But /home/apps/juicer/1.6/CPU/common/ does. And it contains the scripts you mention above:

The same happens for:
stats_sub.awk
collisions.awk
collisions_dups.awk
collisions_dedup_rearrange_cols.awk
juicer_tools
juicer_postprocessing.sh
statistics.pl
check.sh

Also, in the github repo it is stated that if you clone it, the version being downloaded is juicer 2, for 1.6 you have to download the source code. Which I did.

I will take a look at this and see if I can fix it.

[franlat]

Can I see the command that you use to launch Juicer? To see if I can replicate the errors.

[franlat]

Looking at juicer.sh, seems like it needs the directory scripts to be in the root path ($juiceDir). I created that using the Slurm scripts.

In your command, do you use the -D option?

Let's try again, and see if that solves it.

[NoeliaDiaz]

#!/bin/bash
#SBATCH --job-name=juicer
#SBATCH --output=juicer.out
#SBATCH --error=juicer.err
#SBATCH --cpus-per-task=16
#SBATCH --mem=64G
#SBATCH --time=96:00:00
#SBATCH --mail-type=END,FAIL
#SBATCH --mail-user=

set -euo pipefail
set -x

module purge
module load juicer/1.6
module load bwa
module load samtools
module load java/17.0.5

JUICER_CPU="/home/apps/juicer/1.6/CPU"

WORKDIR="/path/to/workingdir/juicer"
TOPDIR="$WORKDIR/juicer_output"
REFDIR="$WORKDIR/references"

GENOME="$REFDIR/GCF_905237075.1_dlabrax2021_genomic.fna"
CHROMSIZES="$REFDIR/genome.chrom.sizes"
SITES="$REFDIR/MboI.txt"

echo "[INFO] TOPDIR=$TOPDIR"
echo "[INFO] Using Juicer CPU path: $JUICER_CPU"

test -d "$TOPDIR" || { echo "[ERR] No existe TOPDIR: $TOPDIR"; exit 2; }
test -f "$GENOME" || { echo "[ERR] No existe GENOME: $GENOME"; exit 2; }
test -f "$CHROMSIZES" || { echo "[ERR] No existe genome.chrom.sizes"; exit 2; }
test -f "$SITES" || { echo "[ERR] No existe MboI.txt"; exit 2; }
test -x "$JUICER_CPU/juicer.sh" || { echo "[ERR] No encuentro juicer.sh en CPU"; exit 2; }

bash "$JUICER_CPU/juicer.sh"
-d "$TOPDIR"
-D "$JUICER_CPU"
-g dlabrax2021
-s MboI
-z "$GENOME"
-p "$CHROMSIZES"
-y "$SITES"
-t $SLURM_CPUS_PER_TASK
-S final

echo "[INFO] Juicer RESUME (final) COMPLETADO EXITOSAMENTE."

I tried Slurm first and then CPU. Saw the files that I need, but in a different directory, as needed/called by juicer, that's why I suggested that maybe the easy fix would be for you to move those files/directories to the needed location.

Thanks for your help with this issue.

[franlat]

Can you try without the -D option? Now it should look for the scripts in SLURM by default.

[NoeliaDiaz]

Hi,
I've tried but got an error because the Juicer inside SLURM looks for helpers in scripts/commons related to -D, and the module in the cluster contains all the needed files in SLURM/scripts/ but not in SLURM/scripts/common/. I am currently trying a fix that consists of temporarily copying those files to a temporary folder with the expected hierarchy (hopefully) and calling juicer.sh this time using -D, pointing to my fix. I also export queue/long_queue to main, trying to avoid the commons/log that does not exist. I sent the job now, I'll keep you posted once I get the result.

[franlat]

Ok! Let me know if it works, and then I can re-organize the original installation. It's a very confusing tool, indeed.