Homec4science
Diffusion scATAC-seq (master) Browse

scATAC-seq/scripts/10xgenomics_PBMC_5k_motifs_classification_2/classification_ctcf_motif.sh
7b3b7bc40180master

Last Change
Raw File

classification_ctcf_motif.sh
View Options

# some paths
## directories
results_dir='results/10xgenomics_PBMC_5k_motifs_classification_2'
data_dir='data/10xgenomics_PBMC_5k_motifs'
## input
file_mat_open="$data_dir/ctcf_motifs_10e-6_open_bin1bp_read_atac.mat"
file_mat_1nucl="$data_dir/ctcf_motifs_10e-6_1nucl_bin1bp_fragment_center.mat"
file_mat_seq="$data_dir/ctcf_motifs_10e-6_sequences.mat"
## file with seeds
file_seed=$results_dir'/ctcf_motifs_10e-6_seed.txt'
mkdir -p $results_dir
touch $file_seed
# parameters
n_iter='20'
n_shift='21'
n_core=24
# open chromatin and nucleosomes
for k in 1 2 3 4 5 6 7 8 9 10
do
seed=$(< /dev/urandom tr -dc _A-Z-a-z-0-9 | head -c${1:-15};echo)
file_prob=$results_dir/'ctcf_motifs_10e-6_open_bin1bp_read_atac_1nucl_bin1bp_fragment_center_'$k'class_prob.mat4d'
file_mod1=$results_dir/'ctcf_motifs_10e-6_open_bin1bp_read_atac_'$k'class_model.mat'
file_mod2=$results_dir/'ctcf_motifs_10e-6_1nucl_bin1bp_fragment_center_'$k'class_model.mat'
file_mod3=$results_dir/'ctcf_motifs_10e-6_sequences_'$k'class_model.mat'
file_aic=$results_dir/'ctcf_motifs_10e-6_open_bin1bp_read_atac_'$k'class_aic.txt'
echo "$file_prob $seed" >> $file_seed
bin/EMJoint --read $file_mat_open --seq $file_mat_seq --class $k --shift $n_shift --flip --iter $n_iter --seed $seed --thread $n_core --out $file_prob
bin/ProbToModel --read $file_mat_open --prob $file_prob --thread $n_core 1> $file_mod1
bin/ProbToModel --read $file_mat_1nucl --prob $file_prob --thread $n_core 1> $file_mod2
bin/ProbToModel --seq $file_mat_seq --prob $file_prob --thread $n_core 1> $file_mod3
done
c4science · Help