Scripts to perform the UMI counting and SSMD score calculation

1.	Installation
	
1.1 Download the file:
1.2 unzip 
1.3 Required dependencies:

(i)	perl packages: IO::Uncompress::Gunzip
(ii)	R: reshape2, plyr

2.	Counting Random Sequence Labels (RSLs)/UMIs
  
  2.1 GuideUMI-count-p0.1.pl

This perl script can be used to count RSLs/UMIs from only one fastq/fastq.gz file at a time.

2.1.1 Usage:

perl GuideUMI-count-p0.1.pl --library lib.csv Sample1_reads.fastq.gz

This will give 2 tables named as Sample1.count Sample1.UMI

2.1.2 Help: perl GuideUMI-count-p0.1.pl --help 

2.1.3 Instructions for file formats

Library file format: 

GuideID,GuideSequence,TargetGene

		AATF_03,GGACCCTGAAGCGGACCCCG,AATF
		AATF_04,GATGAAGGGGAAGATGGGGA,AATF
		AATF_05,CTTCAGATGAGCATTAGCAG,AATF
	
Note:  Any other form of annotation is not allowed.


Input file format: 
Both fastq and fastq.gz files are allowed.
Note: File name should have only one underscore in it. Whatever preceeds underscore in the input file name will be used as outputfile name. Data deposited under PRJEB18436.

2.2 Batch-GuideUMI-count-p0.1.pl

This is a wrapper of the above script that will execute UMI counting for more than 2 input files and merge individual count files.

ALERT: Don't count too many files at the same time (< cpu numbers), because it runs as many processes as your input fastq files and requires a lot of memory.

2.2.1 Usage:

perl Batch-GuideUMI-count-p0.1.pl --library lib.csv --step 12 --fastq Sample1_reads.fq,Samplel2_reads.fastq.gz,Sample3_read.fastq.gz


This script will yield a SampleX.count files and a SampleX.UMI files [X stands for 1,2 or 3] for each of the input files. Additionally, it will merge all the count files into one file named as summary_count.output.raw. Subsequent analysis relies on this merged table.

The --step argument can also be used to only merge count files if UMIs were individually counted on several files using the GuideUMI-count-p0.1.pl script described above (section 2.1).

2.2.2 Help: perl Batch-GuideUMI-count-p0.1.pl –help


3. Data normalization and SSMD Calculation


3.1 IRA-SSMD.R

This is the core R script to calculate the SSMD score for one treatment vs control.

 
3.1.1 Usage: 

Rscript --vanilla IRA-SSMD.R inputfile outputprefix count_threshold (if one wants to filter reads below certain counts)

3.1.2 Example: 

Rscript --vanilla IRA-SSMD.R summary_count.output.raw output_prefix 1 

3.2  IRA-SSMD.sh

This can be used for pairwise comparison of several treatments (i.e. time points) with one control (Day4 vs Day10, Day15, Day20, Day30)

This shell script is a wrapper to parse the big table in order to get pairwise tables (traits against control) to calculate SSMD score. Five arguments must be provided.

To see the help information: 
./IRA-SSMD.sh

3.2.1 Usage:

./IRA-SSMD.sh Rscript Inputfile output_prefix count_threshold (If one wants to filter reads below certain counts)

3.2.2 example

./IRA-SSMD.sh IRA-SSMD.R summary_count.output.raw ABC 1 


4.UMI Binning

4.1
Bin-count-TruncatedUMIs.pl 

This is a perl script for binning CRISPR-Cas9 RSL guides based on the common RSL prefix. 

4.1.1 Usage: 
perl Bin-count-TruncatedUMIs.pl <trunclen> <mincount> <input_countfile.csv>

The script reads the RSL guide counts the from the input file
<input_countfile.csv> and writes to standard output. It then bins
together all RSL guides that have the same first <trunclen> bases and
writes the sum counts of the truncated RSLs to standard output. Only
those RSL guides are considered in the binning that have been observed
at least <mincount> times in at least one of the samples.

4.1.2 Input/Output format:

The input file should be a comma separated text file containing the
following columns

RSL.guide,guide.set,Control,Treatment

The first column contains the guide name and its RSL separated by an
underscore '_'. The second column contains the guide set name. The
last two columns contain the RSL guide counts in two samples (control and treatment). 

Example:

RSL.guide,guide.set,Control,Treatment
AATF_01_AAAAGC,AATF_01,37,2

The output format is the same as input format.