CRISPR-screens_NGS_analysis

Installation

1. In target directory, clone the repository

$ git clone https://github.com/Xander-git/CRISPR-screens_NGS_analysis.git

Quick Sample Analysis using an example dataset

1. From inside the repo's main directory, run matlab from the terminal

CRISPR-screens_NGS_analysis$ matlab

2. Add scripts to matlab path

addpath("./src");

3. (a) Run pipeline on a single sample

NGS_pipeline("example_dataset","sample_name", "trimmomatic",{"1665","1666"});

3. (b) Run pipeline on every sample in a dataset directory using the trimmed reads

NGS_batch_pipeline("example_dataset", "trimmomatic",{"1665","1666"});

---

Pipeline Syntax

Pipeline Run

NGS_pipeline(<dataset_name>, <sample_name>, <read_directory>, {<adapter1>,... <adapterN>});

Use Cases: When each sample within a dataset has different adapters to target, or if batch pipeline gets cut off

Executes the code for step 1-5 on every adapter for a single sample

Batch Run

NGS_batch_pipeline(<dataset_name>, <read_directory>, {<adapter1>,... <adapterN>});`

Use Cases: When an entire dataset uses the same adapters

Executes the pipeline run code for every sample within a dataset directory. In the galaxy dataset folder, you should organize your files with a dataset folder containing each individual sample you plan to analyze.The order of folders should look like this.

galaxy_dataset -> dataset -> samples -> reads_folder & bowtie2_bam -> individual adapter files

---

Function Parameter Definitions

<dataset_name>: String with the name of the target dataset directory within the galaxy_dataset directory

<sample_name>: String with the name of the target sample to search for

<read_directory>: String with the name of the directory to use naive exact matching on. By convention it should be "cutadapt" for untrimmed reads or "trimmomatic" for trimmed reads.

{<adapter1>,... <adapterN>}: A cell array with the name of each adapter of interest in the form of a string or character array.

<adapter>: A string with the name of a single adapter

---

Individual Function Calls

It is recommended that you use the pipeline functions above and just change the starting step in the settings. Using the individual functions is more complex due to the lazy-import implementation. In general, step 1 must always be run first before any others.

Step 1: Import fastqsanger

step1_import_fastq(<dataset_name>, <sample_name>, <read_directory>, {<adapter1>,... <adapterN>});

Imports and counts the reads inside a .fastqsanger or .fastqsanger.gz file. The unique sequences and abundance are then stored in a .mat file to motivate matlab import and export. Should be capable of handling a large range of file sizes. Recommended to use .fastqsanger.gz option to save on memory space.

Step 2: Name Exact Matching

step2_nem(<dataset_name>, <sample_name>, {<adapter1>,... <adapterN>});

Imports the data from the .mat variable file created in step 1 and matches it to the guide rna library specified in the settings.

Step 3: Process Bowtie2

step3_bowtie2(<dataset_name>, <sample_name>, <adapter>);

Processes and generates the counts given by the bowtie2 algorithm from galaxy.

Step 4: Reconcile Counts

step4_reconcile(<dataset_name>, <sample_name>, <adapter>);

Compares the NEM and bowtie2 counts of a specific adapter for a given sample and reports the larger of the two as the final count.

Step 5: Combine Adapters for each Sample

step5_combine_adapters(<dataset_name>, <sample_name>, {<adapter1>,... <adapterN>})

Combines the final count for each adapter onto one table for ease of analysis.

---

Helper Functions

Convert guide library .csv to a .mat table

guidelib_csv2table("<name_of_csv_file>");

Converts a csv file with the guide library into a mat file for the pipeline to read. The column order should be the Guide RNA ID, Guide RNA Sequence, Cutting Score, Guide Position, and Library Design Information. See the included CSV file in guide_library for reference.

Get bowtie2 start positions

get_bowtieStartPos("<fpath_to_bamfile>")

A helper function to get the starting positions from bowtie2 given a bam file. The function returns a matlab table with the bowtie2 reference id, the start position, and the corresponding sequence.

---

NGS_Settings.m

The settings for the NGS_pipeline are stored here. These are vital to the underlying code used. This works similar to class properties.

Params.guide_table_file Default: "Cas9_Opt.mat"

Params.read_filetype Default: ".fastqsanger.gz"

Must either be ".fastqsanger.gz" or ".fastqsanger". The pipeline is capable of reading unzipped or zipped fastqsanger files. By default this is set to the zipped files in an effort to save storage space at the expense of a slight runtime increase

Params.cpu_cores Default: true

Can either be true or a specific number. If true, matlab will check how many available cores are available for use and use the maximum amount during parallel processing. Otherwise a specific number will tell the parallel pool to only use that amount. This is useful if sharing cpu cores with other users.

Params.start_step Default: 0

Tells the pipeline which step to start at. Useful in cases where the pipeline crashes at a certain step.

Params.parpool_timeout Default: 1440

Determines how many minutes the parallel process will run before timing out. If NEM continously times out, it usually means this value is too low. In most cases 1,440 minutes is long enough for NEM to finish.

Params.throw_exception Default: true

Tells the pipeline whether to throw exceptions or continue

Params.post_pipeline_cleanup Default: true

once the pipeline is done, it removes any unzipped files and deletes the generated .mat files in order to save space

Filepath Settings

In general, these should not be changed unless you want to change the directory structure. Essential for pipeline algorithms to function properly. All filepaths should end with a path separator.

Params.galaxy_dir Default:"./galaxy_dataset/"

Filepath to where your galaxy data is stored.

Params.mat_workspace_dir Default:"./src/mat_workspace/"

Params.guide_lib_dir Default:"./guide_library/"

Params.results_dir Default: "./results/"

Params.bam_dir Default: "./bowtie2_bam/"