This is a fork of Arnav Moudgil's Calling Card analysis repository (https://github.com/arnavm/calling_cards), with minor modifications to enable analysis of barcoded self-reporting transposon calling cards. Reference: https://www.biorxiv.org/content/10.1101/2021.04.15.439516

The core of the barcoded calling cards piepline starts with your calling cards sequencing data in fastq or gzipped fastq format. The output of this repository is a set of mapped barcoded self-reporting transposon calling cards in qBED format. The output of this pipeline is used to visualize calling card insertions genome-wide and as input for downstream peak calling.

A running demonstration of the barcoded calling cards pipeline is implemented on Code Ocean: https://codeocean.com/capsule/7988662/tree/v1

Dependencies

The Python scripts in this repository require python3, preferrably a relatively recent version (e.g. ≥ 3.5). In addition, you will need to install the following modules:

• pysam
• numpy
• pandas
• twobitreader
• pybedtools
• scipy

The easiest way to install them is:

• If python3 is the default: pip install pysam numpy pandas twobitreader pybedtools scipy
• Otherwise: pip3 install pysam numpy pandas twobitreader pybedtools scipy

To figure out which version of python is installed by default:

• python -V

The main shell script was designed to be executed from a Linux/Unix based high-performance computing environment using either Slurm or LSF job managers. The software requires several common bioinformatics modules including:

• cutadapt
• umi-tools
• star or (NovoAlign)
• samtools
• bedtools
• java

Getting Started

This is a brief guide to run the barcoded calling cards pipeline.
Within your Linux/Unix based high-performance computing environment, create a new directory. For example, name it CALLINGCARDS.
Within this directory, make 3 more directories: output_and_analysis, raw, and CODE.
Download the contents of this repository to the CODE directory.

If your cluster uses Slurm, append the SLURM_header to bulkRNACallingCards and save this as a shell script.

cat SLURM_header.txt bulkRNACallingCardsBarcodes > bulkRNACallingCardsBarcodes.sh

Alternatively, if your cluster uses LSF, append the LSF_header.

cat LSF_header.txt bulkRNACallingCardsBarcodes > bulkRNACallingCardsBarcodes.sh  

Change the relevant path names in bulkRNACallingCardsBarcodes.sh as appropriate
Move the barcode_safelist.txt and manifest.csv out of the CODE directory into the 'CALLINGCARDS' directory.

Example Analysis

To run the software, download one of our barcoded calling card datasets from the Sequence Read Archive (SRA).

Navigate into the 'raw' directory and download the MYOD1 calling card dataset that was jointly prepared with BRB-seq data.
The SRA number for this dataset is : SRR17863637
The link for this dataset is: https://trace.ncbi.nlm.nih.gov/Traces/sra/?run=SRR17863637
The GEO sample information is: GSM5857636: MYOD1 joint BRB.

One way to download this data is using the SRA toolkit:

ml sratoolkit   
fastq-dump --gzip SRR17863637  
#Rename the file as MYOD1_jointBRB.fastq.gz  
mv SRR17863637.fastq.gz > MYOD1_jointBRB.fastq.gz  

Alternatively, we have provided a sub-sampled dataset in the TUTORIAL folder, named: test.fq.gz
Move this file to your raw directory

The SLURM_header and LSF_header are configured to direct bulkRNACallingCards.sh to analyze line 4 of the manifest (MYOD1_jointBRB.fastq.gz) To analyze the test data instead, change the headers to read line 6.

Executing the main script

From within your 'CALLINGCARDS' directory, run the following command to submit the job.

Using Slurm:

sbatch CODE/bulkRNACallingCardsBarcodes.sh  

Using LSF:

bsub < CODE/bulkRNACallingCardsBarcodes.sh    

Expected Results

This script should generate a file named NN_MYOD1_jointBRB_hg38_map_sort_final.ccf with approximately 587277 lines. Each line corresponds to a genomic insertion of a barcoded SRT.

A full test example including all expected intermediate files is provided in the TUTORIAL folder.
The resulting ccf file (qBED format) can be used as input for downstream peak calling using the mammalian calling cards toolkit: https://gitlab.com/rob.mitra/mammalian_cc_tools/