Munoz-Lopez et al. (2020)

Scripts to call L1 insertions from single-cell ATLAS-seq experiments.

Before starting

Dependencies

cutadapt (tested version: 1.14)
bwa (tested version: 0.7.16a)
Picard tools (tested version: 1.136)
bedtools (tested version: 2.25.0)
seqtk (tested version: 1.0; note that seqtk is only required if subsampling of sequencing data is used - to reduce time of analysis in tests)
GNU grep/awk

Prepare the general configuration file and organize your project folders

Place and edit the .atlas.conf file in your home folder according to your configuration. Prepare directories as follow:

project/
|-- annotations
|-- data
|-- results
`-- scripts

The data folder should contain the .fastq files.

Other requirements

A human reference genome sequence (ex:hg19.fa) and its bwa index (ex: hg19.fa.amb, .ann, .bwt, .fai, .pac, .sa). Their location is indicated in the .atlas.conf file.
adjust the paths of annotations files in the annotations/annotations_sc_5atlas.txt and annotations/annotations_sc_5atlas.txt files.

Procedure

Step 1: For each sequencing run, run the atlas-clustering script:

Starting from a sequencing run, this script:

demultiplex the sequencing reads based on samples
trim the reads and map them on the reference genome provided
cluster sequencing reads and identify potential break points

Example for a 3'-atlas-seq run: The barcode file .bc is a tabular text file with 3 columns (index name, index sequence, sample name). An example is provided in the annotation folder.

cd project/results
../scripts/atlas-clustering_v2.2_forktest_nosoft.sh \
	-d 0 \
	-b ../annotations/3pp.bc \
	../data/R05_INS-203.ATLAS-seq.E6T_E6C1_3prime.fastq

Example for a 5'-atlas-seq run: The barcode file 5pp.bc has 3 columns: index name, index sequence, sample name.

cd project/results
../scripts/atlas-clustering_v2.2_forktest_nosoft.sh \
	-d 0 \
	-f \
	-b ../annotations/5pp.bc \
	../data/R07_INS-208.ATLAS-seq_E6C1_5prime_E6C3_5prime.fastq

Step 2: move all result files into a unique folder

mkdir -p project/results/pooled_single_cells
cp project/results/*_*atlas*/* project/results/pooled_single_cells/

Step 3: run the peak calling and annotation script from the pooled folder

This script calls L1 peaks within large local amplification clusters and annotate them.

cd results/pooled_single_cells
project/scripts/atlas-seq_single_cells_Acount.sh

Step 4: additional filters

This filtering step was added to remove artefactual amplifications in the WGA procedure.

cd results/pooled_single_cells
for file in *.3atlas.v2.2forktest_nosoft.best.insertions.10kb.full.annotated.bed ;
do
	name=$( echo -e $file | awk -F "." '{printf $1}' ) ;

	awk '($1~/^#/) || (($5>=3 && $8+$7>=1) && (($19~/L1HS\|Ta/) || (($19=="." && $21!~/L1PA/) && ($17=="." || $11==0 ) && ($NF<=1))))' ${name}.3atlas.v2.2forktest_nosoft.best.insertions.10kb.full.annotated.bed \
	> ${name}.3atlas.v2.2forktest_nosoft.best.insertions.10kb.full.annotated.filtered.bed;

	cut -f1-6 ${name}.3atlas.v2.2forktest_nosoft.best.insertions.10kb.full.annotated.filtered.bed \
	> ${name}.3atlas.v2.2forktest_nosoft.best.insertions.10kb.filtered.true.bed;
done

retrogenomics/MunozLopez2020