GAD STR expansion detection pipeline

• Author: Anne-Sophie Denommé-Pichon
• Version: 0.0.1
• Licence: AGPLv3
• Description: this pipeline allows to get STR genotype from short-read genomes on the locus specified. It uses ExpansionHunter, Tredparse and GangSTR. It computes genotypes called by the tools and identifies STR expansions using 3 outlier detection methods to highlight abnormal repeat counts.

Setup

• Fill the configuration file config.sh. There is an example in the repository.
• Create samples.list (bam file names without .bam). There is an example in the repository.

Usage

For now, scripts have to be launched from the clone directory.

Calling STRs

Launch launch_pipeline.sh:

nohup ./launch_pipeline.sh samples.list &

Dependencies:

• config.sh
• samples.list
• pipeline.sh
• wrapper_delete.sh
• wrapper_ehdn.sh
• wrapper_expansionhunter.sh
• wrapper_gangstr.sh
• wrapper_transfer.sh
• wrapper_tredparse.sh

Identifying outliers

To highlight abnormal repeat counts, the pipeline identified outliers using 3 methods:

1. repeats counts at a given locus > normal (in the gray zone or pathological zone)
2. repeats counts at a given locus > 99th percentile or
3. repeats counts at a given locus ≥ 4 standard deviations above the mean (Z-score ≥ 4).

Launch launch_results.sh:

nohup ./launch_results.sh samples.list &

Dependencies:

• config.sh
• samples.list
• patho.csv
• getResults.py
• launch_str_outliers.sh
• str_outliers.py

Future work

Another tool, ExpansionHunter DeNovo, will be added in the pipeline.