Alternative polyadenylation

Alternative polyadenylation#

This mini-protocol converts aligned RNA-seq reads into an analysis-ready alternative-polyadenylation (APA) phenotype matrix for apaQTL analysis.

Miniprotocol Timing#

Timing: TBD

Overview#

APA calling derives 3′UTR regions, converts transcriptome BAM files to per-base coverage, and runs DaPars2 to estimate percentage of distal polyadenylation site usage (PDUI). APA imputation and QC imputes missing PDUI values, quantile-normalizes the matrix, and can replace internal sample identifiers with analysis identifiers.

The full calling route requires transcriptome BAM files, which are not included with the example data. A precomputed chromosome-22 DaPars result is provided so that the imputation and renaming routes can be run independently.

Steps#

Choose the route that matches the available starting data; the six commands are not always one mandatory chain.

Analysis goal

Commands to run, in order

Inputs

Call and process APA from transcriptome alignments

1 → 2 → 3 → 4 → 5

input/reference_data/protocol_example.apa.chr22.gtf; transcriptome BAM files under output/rnaseq/bam/ (not bundled)

Resume from per-sample coverage files

3 → 4 → 5

output/apa/wig/*.wig; matching depth files; output/apa/protocol_example.apa.chr22_3UTR.bed

Impute and normalize an existing chromosome-22 DaPars result

5

output/apa/apa_chr22/Dapars_result_result_temp.chr22.txt

Impute, normalize and rename samples

5 → 6

DaPars result above; input/covariate/protocol_example.apa_matchtable.txt

Within the selected route, run the listed commands in numerical order.

1. Build the 3′UTR reference#

What it does: UTR_reference extracts transcript-level 3′UTR intervals from the reference GTF for DaPars2.

sos run pipeline/apa_calling.ipynb UTR_reference \
  --cwd output/apa \
  --hg-gtf input/reference_data/protocol_example.apa.chr22.gtf

2. Generate coverage and read-depth files#

What it does: bam2tools converts each transcriptome BAM into a per-base WIG file and a matching read-depth summary.

sos run pipeline/apa_calling.ipynb bam2tools \
    --cwd output/apa \
    --bam-dir output/rnaseq/bam

3. Build the DaPars2 configuration#

What it does: APAconfig creates the sample mapping and DaPars2 configuration files from the coverage directory and 3′UTR annotation.

sos run pipeline/apa_calling.ipynb APAconfig \
  --cwd output/apa \
  --bfile output/apa/wig \
  --annotation output/apa/protocol_example.apa.chr22_3UTR.bed

4. Estimate chromosome-level PDUI#

What it does: APAmain runs DaPars2 for chromosome 22 and writes the unprocessed PDUI result used by the QC module.

sos run pipeline/apa_calling.ipynb APAmain \
    --cwd output/apa \
    --chrlist chr22 \
    --chr-prefix true \
    --dapars-path code/SoS/molecular_phenotypes/calling/apa

5. Impute and normalize PDUI#

What it does: APAimpute imputes missing chromosome-level PDUI values and applies quantile normalization.

sos run pipeline/apa_impute.ipynb APAimpute \
    --cwd output/apa \
    --chrlist chr22

6. Rename samples and assemble the final matrix#

What it does: APArename maps sample identifiers, combines requested chromosomes, and bgzip-indexes the final APA phenotype matrix.

sos run pipeline/apa_impute.ipynb APArename \
    --cwd output/apa \
    --chrlist chr22 \
    --match input/covariate/protocol_example.apa_matchtable.txt

Output Files#

Step

Output filename and relative path

Description

1

output/apa/protocol_example.apa.chr22.bed; output/apa/protocol_example.apa.chr22.transcript_to_geneName.txt; output/apa/protocol_example.apa.chr22_3UTR.bed

BED12 annotation, transcript-to-gene map and DaPars2 3′UTR intervals

2

output/apa/*.wig; output/apa/*.depth

Per-sample coverage and read-depth files

3

output/apa/sample_mapping_files.txt; output/apa/sample_configuration_file.txt

DaPars2 sample mapping and configuration

4

output/apa/apa_chr22/Dapars_result_result_temp.chr22.txt

Raw chromosome-22 DaPars2 PDUI result

5

output/apa/apa_chr22/Dapars_result_impute_chr22.bed

Imputed and normalized chromosome-22 PDUI matrix

6

output/apa/apa_chr22/Dapars_result_impute_renamed_chr22.bed; output/apa/apa_chr22/Dapars_result_impute_renamed_chr22.bed.gz; output/apa/apa_chr22/Dapars_result_impute_renamed_chr22.bed.gz.tbi; output/apa/Dapars_allchrom_renamed.bed; output/apa/Dapars_allchrom_renamed.bed.gz; output/apa/Dapars_allchrom_renamed.bed.gz.tbi

Renamed chromosome-level and combined analysis-ready APA phenotype matrices

Anticipated Results#

The complete calling route produces 3′UTR intervals, per-sample coverage summaries, a DaPars2 configuration and one raw PDUI result per requested chromosome. Step 5 produces an imputed chromosome-level phenotype matrix. Step 6 is optional but recommended when the BAM identifiers differ from the identifiers used by covariates and genotypes; its combined bgzip-compressed BED file is the input for downstream apaQTL association testing.

The bundled example starts at step 5 because transcriptome BAM files are not included.

Command interface#

sos run pipeline/apa_calling.ipynb -h

sos run pipeline/apa_impute.ipynb -h