IntGenomicsLab · AmberVerhasselt · Sep 14, 2026 · Sep 2, 2026 · Sep 2, 2026 · Sep 3, 2026
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -8,12 +8,14 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ### `Added`
 
 - [#190](https://github.com/IntGenomicsLab/lrsomatic/pull/190) - Added mutational signature analysis: SigProfilerMatrixGenerator matrices (SBS/DBS/ID) and per-sample COSMIC signature fitting with SigProfilerAssignment on the phased somatic VCF, with CHM13-T2T support via a purpose-built image (`ghcr.io/ljwharbers/sigprofiler`) and CHM13-renormalised COSMIC signatures; new `--sigprofiler_*` / `--download_sigprofiler_genome` / `--skip_signatures` parameters (@ljwharbers).
+- [#188](https://github.com/IntGenomicsLab/lrsomatic/pull/188) - Added `modkit_args` (default `--cpg --modified-bases 5mC`) to control the arguments passed to `modkit pileup`, and `modkit_phased` to run the pileup with `--phased` on the Longphase-haplotagged BAMs, producing `_hp1`, `_hp2` and `_combined` bedMethyl files per sample (@ljwharbers).
 - [#176](https://github.com/IntGenomicsLab/lrsomatic/pull/176) - Added `LRSOMATICREPORT` as the final pipeline step: a self-contained per-sample HTML report covering small variants, structural variants, copy number and QC. Skip it with `--skip_report`; choose the gene panel selected on load with `--report_gene_panel` (@ljwharbers).
 - [#176](https://github.com/IntGenomicsLab/lrsomatic/pull/176) - Vendored the [lrsomatic_report](https://github.com/ljwharbers/lrsomatic_report) v1.3.2 tool source at `assets/lrsomatic_report`, so `nextflow run IntGenomicsLab/lrsomatic` ships it without a submodule checkout (@ljwharbers).
 - [#176](https://github.com/IntGenomicsLab/lrsomatic/pull/176) - Added a `solution_dirs` output to the WAKHAN module so its per-solution copy-number plots can be staged downstream (@ljwharbers).
 
 ### `Changed`
 
+- [#188](https://github.com/IntGenomicsLab/lrsomatic/pull/188) - `MODKIT_PILEUP` now runs `modkit pileup` with `--cpg --modified-bases 5mC` by default; previously it ran with no arguments, and that unfiltered pileup (every modification code at every sequence context) produced 30-45 GB per sample. Reruns therefore give different bedMethyl content: only 5mC rows at CpG sites, so 5hmC and 6mA calls and non-CpG positions are no longer reported. Set `--modkit_args=''` (the `=` form; `--modkit_args ''` reaches the pipeline as `true` and is rejected by parameter validation on Nextflow 25) or an empty `modkit_args` entry in a params file to get the previous unfiltered output (@ljwharbers).
 - [#186](https://github.com/IntGenomicsLab/lrsomatic/pull/186) - Re-synced the vendored [lrsomatic_report](https://github.com/ljwharbers/lrsomatic_report) to v1.3.0, which adds tickbox dropdown filters on the categorical columns of both variant tables and turns the report's gene panel selector into checkboxes (@ljwharbers).
 - [#176](https://github.com/IntGenomicsLab/lrsomatic/pull/176) - Re-synced the vendored [lrsomatic_report](https://github.com/ljwharbers/lrsomatic_report) to v1.3.2: facet dropdown counts follow the active filters, opening a facet menu no longer resets the table's horizontal scroll, a flatter clinical theme, and inline code comments trimmed to one line (@ljwharbers).
 - [#186](https://github.com/IntGenomicsLab/lrsomatic/pull/186) - `--report_gene_panel` now takes a comma-separated list, so several panels can be applied at once: a variant or SV is kept if it hits any of them. Panel values are also validated at launch instead of failing inside the report task (@ljwharbers).
@@ -26,6 +28,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### `Fixed`
 
+- [#188](https://github.com/IntGenomicsLab/lrsomatic/pull/188) - `MODKIT_PILEUP` now runs a patched modkit 0.6.4 built from [ljwharbers/modkit@pacbio-conflict-fix](https://github.com/ljwharbers/modkit/tree/pacbio-conflict-fix): `ghcr.io/ljwharbers/modkit:0.6.4-pacbiofix-6e0afa2` under Docker, and the native SIF `oras://ghcr.io/ljwharbers/modkit-sif:0.6.4-pacbiofix-6e0afa2` under Singularity/Apptainer (unless `singularity_pull_docker_container` is set, which falls back to the Docker image). It keeps reads whose 5mC and 5hmC probabilities sum above 1, fixes pileup on PacBio-style MM tags, and honours `--phased` and `--modified-bases` in the general pileup workers that modkit uses for PacBio BAMs with 6mA calls (stock modkit wrote empty `_hp1`/`_hp2` files and an `h` row next to every `m` row for those). Stock modkit 0.4.3-0.6.4 silently dropped 32-65 % of reads from recent PacBio HiFi BAMs (Jasmine >= 26.1.3) and returned empty `--cpg` pileups ([nanoporetech/modkit#612](https://github.com/nanoporetech/modkit/issues/612); fixes proposed upstream in [nanoporetech/modkit#720](https://github.com/nanoporetech/modkit/pull/720)). The image is `linux/amd64` only and Conda is not supported for this step: `MODKIT_PILEUP` stops with an error under `-profile conda`/`mamba`, use `--skip_modkit` there. The module should return to the nf-core biocontainer once a modkit release includes the fix (@ljwharbers).
 - [#181](https://github.com/IntGenomicsLab/lrsomatic/pull/181) - NanoPlot pre- and post-alignment statistics now reach MultiQC. `ch_nanoplot_pre_txt` and `ch_nanoplot_post_txt` were declared empty and mixed into the MultiQC inputs, but never assigned from `NANOPLOT_PRE.out.txt` / `NANOPLOT_POST.out.txt`, so the NanoStat section was silently missing from every report (@ljwharbers).
 - [#181](https://github.com/IntGenomicsLab/lrsomatic/pull/181) - `NANOPLOT_PRE` now includes the replicate in its output prefix, so samples with more than one replicate no longer collapse into a single MultiQC sample (@ljwharbers).
 - [#181](https://github.com/IntGenomicsLab/lrsomatic/pull/181) - samtools stats/flagstat/idxstats and mosdepth outputs are now prefixed `<sample>_<type>` instead of `<sample>`. The tumor and normal files of a matched pair shared a basename, so MultiQC logged `Duplicate sample name found! Overwriting` and reported only one of the two; the general statistics table now has one row per BAM (`<sample>_tumor`, `<sample>_normal`), which the post-alignment NanoStat row joins (@ljwharbers).

diff --git a/conf/modules.config b/conf/modules.config
@@ -330,6 +330,14 @@ process {
     //
 
     withName: '.*:MODKIT_PILEUP' {
+        ext.args = {
+                [
+                    // Nextflow turns `--modkit_args ''` (and a bare `--modkit_args`) into `true`: a boolean that nf-schema rejects on
+                    // Nextflow 25, the string 'true' on Nextflow 26. The documented form is `--modkit_args=''`; this backstop treats `true` as "no arguments"
+                    (params.modkit_args == null || params.modkit_args in [true, 'true']) ? '' : params.modkit_args,
+                    params.modkit_phased ? '--phased' : ''
+                ].join(' ').trim()
+            }
         publishDir = [
             path: { "${params.outdir}/${meta.id}/methylation/${meta.type}/modkit_pileup" },
             mode: params.publish_dir_mode,

diff --git a/docs/output.md b/docs/output.md
@@ -231,15 +231,20 @@ QC outputs are placed under `tumor/` for all samples, and additionally under `no
 ├── methylation
 │   ├── tumor
 │   │   └── modkit_pileup
-│   │       └── sample.bed.gz
+│   │       ├── sample.bed.gz           # default
+│   │       ├── sample_hp1.bed.gz       # --modkit_phased only
+│   │       ├── sample_hp2.bed.gz       # --modkit_phased only
+│   │       └── sample_combined.bed.gz  # --modkit_phased only
 │   ├── normal                          # paired samples only
 │   │   └── modkit_pileup
-│   │       └── sample.bed.gz
+│   │       └── ...                     # same layout as tumor
 ```
 
-| File                                         | Description                                                                         |
-| -------------------------------------------- | ----------------------------------------------------------------------------------- |
-| `{tumor,normal}/modkit_pileup/sample.bed.gz` | Modkit pileup BED file containing per-CpG methylation frequency and coverage values |
+| File                                                   | Description                                                                                                                                                                                                     |
+| ------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| `{tumor,normal}/modkit_pileup/sample.bed.gz`           | Modkit pileup bedMethyl table (bgzip) with per-strand methylation frequency and coverage. By default (`--modkit_args '--cpg --modified-bases 5mC'`) rows are 5mC calls at CpG sites only.                       |
+| `{tumor,normal}/modkit_pileup/sample_{hp1,hp2}.bed.gz` | With `--modkit_phased`: bedMethyl tables restricted to reads carrying haplotype tag `HP:1` or `HP:2` from the Longphase-haplotagged BAM.                                                                        |
+| `{tumor,normal}/modkit_pileup/sample_combined.bed.gz`  | With `--modkit_phased`: bedMethyl table over all reads, including untagged ones (equivalent to the unphased default output). There is no separate file for untagged reads; they only contribute to `_combined`. |
 
 </details>