predict_model.sh

##predict all cpg
mkdir $WD/$5
cd $WD/$5
Rscript $WD/src/normalize_from_template.R $WD/$6/data_final $WD/$7/data_norm_info data_final_norm
#generate all data (modify 4-20-20)
wc -l $WD/$8/cpg_no_chrM >num_cpg_info.txt
numberCpG=`cat num_cpg_info.txt |cut -d ' ' -f1`
python3 $WD/src/generate_data_array.py all_data data_final_norm $WD/$8/cpg_no_chrM 1 $numberCpG 45

#pred
#use the best epoch from rnn and cnn and joint module for below prediction
cp $WD/$1/$2* .
cp $WD/$1/$3* .
cp $WD/$1/$4* .

python3 -u $WD/src/continue_training_pred.py all_label1 all_label2 $4 $WD/$1/train_small.npz all_data.npz $2 $3 $WD/$9/genome_encoding.npz

cut -f2 all_label1 >all_label1_cut
cut -f2 all_label2 >all_label2_cut
paste all_label1_cut all_label2_cut >all_pred
paste $WD/$6/data_final_coord all_pred >all_coord_pred_hmc_mc
#all_coord_pred_hmc_mc: coord, hmc, mc

#adjust prediction by mlml
Rscript $WD/src/adjust_prediction.R all_coord_pred_hmc_mc pred_hmc pred_mc
python3 $WD/src/format_as_mlml.py pred_mc pred_mc_format
python3 $WD/src/format_as_mlml.py pred_hmc pred_hmc_format
$WD/MethPipe/bin/mlml -v -m pred_mc_format -h pred_hmc_format  -o data_mlml
cut -f1-5 data_mlml >all_coord_pred_mc_hmc_final
#all_coord_pred_mc_hmc_final: coord, mc, hmc