search

precimed / mixer

Causal Mixture Model for GWAS summary statistics
GNU General Public License v3.0
65 stars 18 forks source link

Error when running ./mixer.py fit1 #46

Closed FernandaSantos13 closed 3 years ago

FernandaSantos13 commented 3 years ago

Hi Alex,

I am trying to run MiXer 1.3 again for my data and I am getting some errors:

####### Code module purge # clear any inherited modules module load devel/CMake/3.15.3-GCCcore-8.3.0 module load devel/Boost/1.74.0-GCC-10.2.0 module load lang/Python/3.7.4-Anaconda3-2021.03

Univariate model

python3 $SCRATCH/fsantos/mixer_v1-3_20210429/mixer/precimed/mixer.py fit1 \ --trait1-file $INPUT_PATH/PGC_MDD_2018_sumstats_mixer_qc_noMHC.csv.gz \ --out $OUTPUT_PATH/PGC_MDD_2018_sumstats_mixer_qc_noMHC_vPAPER_May2021.fit.rep${SLURM_ARRAY_TASK_ID} \ --extract $REF_PATH/1000G_EUR_Phase3_plink/1000G.EUR.QC.prune_maf0p05_rand2M_r2p8.rep${SLURM_ARRAY_TASK_ID}.snps \ --bim-file $REF_PATH/1000G_EUR_Phase3_plink/1000G.EUR.QC.@.bim \ --ld-file $REF_PATH/1000G_EUR_Phase3_plink/1000G.EUR.QC.@.run4.ld \ --lib $SCRATCH/fsantos/mixer_v1-3_20210429/mixer/src/build/lib/libbgmg.so \

########

######## .out file

$ cat slurm-37647.out

The following have been reloaded with a version change: 1) compiler/GCCcore/8.3.0 => compiler/GCCcore/10.2.0 2) lib/zlib/1.2.11-GCCcore-8.3.0 => lib/zlib/1.2.11-GCCcore-10.2.0 3) tools/bzip2/1.0.8-GCCcore-8.3.0 => tools/bzip2/1.0.8-GCCcore-10.2.0

INFO:root:init(lib_name=/genome/scratch/diniz/fsantos/mixer_v1-3_20210429/mixer/src/build/lib/libbgmg.so, context_id=0) INFO:root:init_log(/genome/scratch/diniz/fsantos/DINIZprojects/AD_vs_BP/mixer_analyses/PGC_MDD_2018_sumstats_mixer_qc_noMHC_vPAPER_May2021.fit.rep.log) INFO:root:log_message(***

Fernanda

ofrei commented 3 years ago

@FernandaSantos13 there should be an .log file with more info, could you attach it here or send it to my e-mail ?

FernandaSantos13 commented 3 years ago

@ofrei Yes! For sure! PGC_MDD_2018_sumstats_mixer_qc_noMHC_vPAPER_May2021.fit.rep.log

ofrei commented 3 years ago

This part is relevant, it shows that after applying --extract argument the number of tag SNPs dropped to 0. This is most likely because --extract file does not exist. It ends with rep.snps while it should be rep1.snps, rep2.snps, etc - dependig on SLURM_ARRAY_TASK_ID. Do you have --array in your SLURM job? Take a look at https://github.com/precimed/mixer#upgrade-nodes-from-mixer-v12-to-v13

 constrain analysis to 8003224 tag variants (due to trait1_file='/genome/scratch/diniz/fsantos/DINIZprojects/AD_vs_BP/mixer_analyses/PGC_MDD_2018_sumstats_mixer_qc_noMHC.csv.gz')
20210520 14:05:54.397924     constrain analysis to 0 tag variants (due to extract='/genome/scratch/diniz/fsantos/DINIZprojects/reference_files/mixer1.3_20210430/1000G_EUR_Phase3_plink/1000G.EUR.QC.prune_maf0p05_rand2M_r2p8.rep.snps')
20210520 14:05:54.416866
FernandaSantos13 commented 3 years ago

@ofrei No... I did not have this tag for fit 1. Thank you. It is running again with the following flags:

Job name

SBATCH --job-name=mixer_MDD_fit1_20May2021

Job resources

Wall clock limit:

SBATCH --time=0-24:00:00

SBATCH --cpus-per-task=8

Max memory usage:

SBATCH --mem-per-cpu=7600M

SBATCH --array=1-20

I will let you know if it worked.

FernandaSantos13 commented 3 years ago

It worked!!! Thanks!