Script Style Guide

Coding style

• Comments are preceded by 1 #.
• Where scripts are organised into sections, section headers are all CAPITALS

Script naming conventions

• Job submission scripts to be numbered. Numbers should be sequential even if tasks aren't necessarily.
• Job submission scripts that "start" with batch, indicate an array job with some element of parallelisation across samples (not necessarily 1 job per sample).
• Processing scripts not numbered

Job submission script format

• SLURM scheduler
• Start with a series of checks to confirm variables are as expected (e.g. command line flags) and source the config file:

## print start date and time
echo Job started on:
date -u

export PROJECT=$1
source ./ATACSeq/config/config.txt 

##check array specified and exit if not
if [[ ${SLURM_ARRAY_TASK_ID} == '' ]]
then 
    { echo "Job does not appear to be an array. Please specify --array on the command line." ; exit 1; }
fi

## check step method matches required options and exit if not
if [[ ! $2 =~ "FASTQC" ]] && [[ ! $2 =~ "TRIM" ]] && [[ ! $2 =~ "ALIGN" ]] && [[ ! $2 =~ "ENCODE" ]] &&[[ ! $2 == '' ]];
then 
    { echo "Unknown step specified. Please use FASTQC, TRIM, ALIGN, ENCODE or some combination of this as a single string (i.e. FASTQC,TRIM)" ; exit 1; }            
fi

• Use dashed line to separate checks from main body

#-----------------------------------------------------------------------#

• End with completion statement:

echo 'EXIT CODE: ' $?

Processing script format

• Specify required variables and inputs in header:

## <single line description>

## REQUIRES the following variables from config file

# <comma separated list>

## REQUIRES the following software

# <comma separated list>

## INPUT

# <description of file>

## OUTPUT

# *
## (command line) VARIABLES

#-------------------------------#

MAIN

where MAIN is the code body of the script.

• End with a completion statement

if [[ $? == 0 ]] # if exit code is 0
	then echo "Alignment and post filtering complete"
	date -u
fi