This is a guide to run the K-mer pipeline and get models going

• before running the pipeline you need to

I. clone kmer pipeline in github

https://github.com/Thilanka-lt/pCRE_identification_mod.git

II. Activate environments

conda create -n pCR_identification python=3.6.4
conda activate pCR_identification
pip install scikit-learn==0.23.1
pip install imbalanced-learn==0.7.0
pip install matplotlib==3.3.2
pip install pandas==1.1.5
pip install numpy==1.18.5
pip install statsmodels==0.12.1
pip install joblib==0.17.0
pip install scipy==1.5.4

• Once you set up your environment you can activate environment in a shell script using : conda activate path to environment/pCR_identification

Ex:

conda activate /mnt/home/ranawee1/anaconda3/envs/pCR_identification

• You can install required packages for the virtual environment using the requirement.txt file in the repo

conda create -n pCR_identification python=3.6.4
conda activate sc_dim_red
pip install -r requirement.txt

1. Set Up Your Files:

1.a. Inside directory where your model is going to run make sub directories for FASTA files and Motif Lists:

mkdir FastaFiles
mkdir MotifLists

1.b. Copy/generate fasta file (file that includes sequence information) for all the genes including the region of interest to FastaFiles

python /mnt/gs21/scratch/ranawee1/05_Singel_cell/scripts/00_extract_regions_from_gff.py -g TAIR10_GFF3_genes.gff -c TAIR10_chr_all.fas -u 1000 -d 0 -r 3 -i 500

• you can also copy from

/mnt/gs21/scratch/ranawee1/05_Singel_cell/data/TAIR10_chr_all_upstream_1000_ingene_500_from_TSS_downstream_0.fas

1.c. copy your negative and positive examples files to FastaFiles directory

• These are text files with gene names

Ex:

AT1G01070
AT1G01140
AT1G01230
AT1G01440
AT1G01470
AT1G01550
AT1G01620

1.d Make singleton and paired kmer list. Reverse complement sequences are separated by “.”, pairs separated by a space (k = length of kmer you want):

• First navigate to MotifLists directory (you can dedicate a directory to save all the kmer lists and call that directory using -path_m flag once you are running the pipeline)
• Then,

python ./Pairwise_kmers.py -f make_pairs2 –k 6

• change the -k flag to generate kmers of preferred length

2. Running the model:

• run this code in the base directory

python /mnt/gs21/scratch/ranawee1/Maddy_Rajneesh_prj/Motif_dicovary/pCRE_identification/000_pipeline_for_MotifDiscovery_stratified_for_test.py -pos_ID positive_set.txt -neg_ID negative_set.txt -fasta TAIR10_chr_all.fas_upstream_1000_ingene_500_downstream_0.fas -bal n -holdout 0.2 -kmers 5mers_withRC.txt,6mers_withRC.txt,7mers_withRC.txt -pval 0.05 -smote y -stratified_for_test n -path_p /mnt/gs21/scratch/ranawee1/Maddy_Rajneesh_prj/Motif_dicovary/pCRE_identification -path_k /mnt/gs21/scratch/ranawee1/Maddy_Rajneesh_prj/Motif_dicovary/models/MotifLists -short n