DMS-Seq tutorial

1. Clone this repository

git clone https://github.com/goodarzilab/DMS-tutorial.git

2. Install the conda environment

conda env create -f DMS-Seq.yml

3. Clone and Install RNA-Framework (https://rnaframework-docs.readthedocs.io/en/latest/)

To clone: git clone https://github.com/dincarnato/RNAFramework

Add rnaframework scripts to path variable (replace /path/to/RNAFramework with the absolute path of the cloned RNAFramework folder)

export PATH=$PATH:/path/to/RNAFramework

4. Trim the reads

Assumes that the fastqs are in the same folder as the trim_galore.sh script. If needed copy over the script to the fastq folder

bash trim_galore.sh

5. Create a bowtie reference for the sequence

Prepare a fasta (.fa file) for the sequence to align against. For example, the file MYC_5UTR.fais included in the repository for formatting purposes. Put it in a folder called /ref and run the following commands in the /ref folder.

Make a bowtie reference for the fasta file using the command bowtie2-build -f {reference.fa} {output_ref_name} eg: bowtie2-build MYC_5utr.fa MYC_5utr

Now ref folder will contain the .fa file, and all the other files associated with the Bowtie reference, with different extensions but containing the suffix {output_ref_name}

6. Align the trimmed fastqs

Trimmed fastqs should now have the name *val_{1/2}.fq.gz. Similar to trim_galore.sh script, the align.sh script assumes that the trimmed fastqs are in the same folder.

bash align.sh -ref {/path/to/refFolder/with/Bowtie_suffix}

eg: bash align.sh -ref ./ref/MYC_5utr

7. Process bams

Folder should now contain bam files (aligned), now we use rnaframework to count mutations, normalize and predict structure.

To count mutations:

• rf-count -r -m -f ref/reference.fa sample.bam -o output_folder. To process multiple replicates at once you can use the * (wildcard character).
• eg: rf-count -r -m -f ref/MYC_5utr.fa siRBM42*.bam -o output_folder

To normalize (individual replicates):

• rf-norm -t output_folder/*.rc -i output_folder/index.rci -sm 4 -nm 2 -rb AC

To combine replicates into one xml file:

• rf-combine Rep1_norm/{refrence_name}.xml Rep2_norm/{refrence_name}.xml -o {combined_output_folder}
• eg: rf-combine DMS-siRBM42_Rep1_S22.srt_norm/MYC_5UTR.xml DMS-siRBM42_S21.srt_norm/MYC_5UTR.xml -o DMS-siRBM42_combined

To do structure prediction:

• rf-fold -g -ct --folding-method 2 {combined_output_folder}/{file}.xml -dp -o {fold_output}
• eg: rf-fold -g -ct --folding-method 2 DMS-siRBM42_combined/MYC_5UTR.xml -dp -o DMS_siRBM42_fold

The fold command outputs 3 folders, dotplot (for IGV visualization), images (for linear folding plot) and structures for .ct file

8. Inspect reactivites.

The two *.html notebook files in the repository contains R and python code to parse through the xml files to read structures and create outputs that qc reactivities between replicates and create visualizations