Skip to content

About

With this workflow you can use the output of MPRAsnakeflow (reporter experiment barcode file) together with the MPRA library sequence design file to quantify the MPRA reporter expression. It can quantify variants as well as elements. using BCalm.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Repository files navigation

IGVF MPRA quantification workflow

With this workflow you can use the output of MPRAsnakeflow (reporter experiment barcode file) together with the MPRA library sequence design file to quantify MPRA reporter expression. It can quantify variants as well as elements. For elements, a label file is needed and the control and test label IDs of oligos. For example, a label file looks like:

shuffled_oligo_1	negative
shuffled_oligo_2	negative
test_oligo_1	test
test_oligo_2	test
test_oligo_3	test

Further, it can quantify using all available barcodes per oligo (config level=barcode) or it can use an aggregated version (config level=oligo). For quantification and statistics it uses BCalm.

Keukeleire, Pia, Jonathan D. Rosen, Angelina Göbel-Knapp, Kilian Salomon, Max Schubach, and Martin Kircher.
"Using individual barcodes to increase quantification power of massively parallel reporter assays."
BMC Bioinformatics 26, no. 1 (2025): 52.

It generates reporter variants, reporter genomic variants, reporter elements, and reporter genomic elements file formats (defined by the IGVF MPRA FG).

Requirements

  • snakemake
  • Apptainer/Singularity

Container images

The mpralib and BCalm images are built and pushed independently to Docker Hub by GitHub Actions. Each workflow runs when its version file, Dockerfile, or build workflow changes on master or development, and can also be started manually. Existing version tags are not overwritten.

  • mpralib: visze/mpralib:<version>, configured in version.mpralib.txt.
  • BCalm: visze/bcalm:<version>, configured in version.bcalm.txt, starting with 0.100.1 (upstream release v0.100.1).

Both images install packages from conda-forge and Bioconda using Pixi. The BCalm quantification rules use the versioned image rather than latest. To upgrade BCalm, update version.bcalm.txt and the r-bcalm pin in workflow/envs/bcalm.yaml together.

Publishing uses the repository variable DOCKERHUB_USERNAME and secret DOCKERHUB_TOKEN, as in the mpralib workflow.

Variant quantification

Input files are:

  • /path/to/data/reporter_experiment_barcode.tsv.gz: contains the counts per barcode.
  • /path/to/data/mpra_library_design.tsv.gz: contains the MPRA library sequence design file.

Example using barcode-level quantification:

conda:

snakemake --sdm conda \
--config level=barcode \
id=test \
count_file=/path/to/data/reporter_experiment_barcode.tsv.gz \
sequence_design_file=/path/to/data/mpra_library_design.tsv.gz \
-c 1 all_variants

apptainer:

snakemake --sdm apptainer \
--config level=barcode \
id=test \
count_file=/path/to/data/reporter_experiment_barcode.tsv.gz \
sequence_design_file=/path/to/data/mpra_library_design.tsv.gz \
--apptainer-args "-B /path/to/data -B $HOME/.cache/snakemake" \
-c 1 all_variants

Here we use an ID test so that the files are tagged with test. Output files are here:

  • results/test/reporter_genomic_variants/test.reporter_genomic_variants.barcode.tsv.gz
  • results/test/reporter_variants/test.reporter_variants.barcode.tsv.gz

With a volcano plot:

  • results/test/quantification/test.barcode.variant.volcano.png

The config option level=barcode can be removed because it is the default.

For running on aggregated counts you can just use level=oligo.

Element quantification

Input files are:

  • /path/to/data/reporter_experiment_barcode.tsv.gz: contains the counts per barcode.
  • /path/to/data/mpra_library_design.tsv.gz: contains the MPRA library sequence design file.

Example using barcode-level quantification:

conda:

snakemake --sdm conda \
--config level=barcode \
id=test \
count_file=/path/to/data/reporter_experiment_barcode.tsv.gz \
sequence_design_file=/path/to/data/mpra_library_design.tsv.gz \
test_label=test \
control_label=negative \
label_file=/path/to/data/labels.tsv.gz \
-c 1 all_elements

apptainer:

snakemake --sdm apptainer \
--config level=barcode \
id=test \
count_file=/path/to/data/reporter_experiment_barcode.tsv.gz \
sequence_design_file=/path/to/data/mpra_library_design.tsv.gz \
test_label=test \
control_label=negative \
label_file=/path/to/data/labels.tsv.gz \
--apptainer-args "-B /path/to/data -B $HOME/.cache/snakemake" \
-c 1 all_elements

Here we use an ID test so that the files are tagged with test. Output files are here:

  • results/test/reporter_genomic_elements/test.reporter_genomic_elements.barcode.tsv.gz
  • results/test/reporter_elements/test.reporter_elements.barcode.tsv.gz

With a density and a volcano plot:

  • results/test/quantification/test.barcode.element.density.png
  • results/test/quantification/test.barcode.element.volcano.png

The config option level=barcode can be removed because it is the default.

For running on aggregated counts you can just use level=oligo.

About

With this workflow you can use the output of MPRAsnakeflow (reporter experiment barcode file) together with the MPRA library sequence design file to quantify the MPRA reporter expression. It can quantify variants as well as elements. using BCalm.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages