and if it's the latter, do you just run WES data as WTGS?

and i suppose an extension to the question is how is it recommended to run the combination of tumor & normal WES + WTS data

( i tried reading thru prior threads here, but couldn't discern a clear answer to what to do without knowing the probe set up front )

I'm attempting to run oncoanalyser in AWS Healthomics; I've specified S3 paths of all the reference files and i've set ignore_igenomes to "true" but it still tries to connect to igenomes and fails, preventing the pipeline from running. Any suggestions here?

i pointed it at an existing s3 location that doesnt' actually contain the igenome ref but is readable by the pipeline and it is proceeding - probably a bug worth looking at - should I report on the github issue tracker?

❌ oncoanalyser/test_full failed 😿 Run Name distraught_mclean Duration 1h 15m 59s Status ❌ Failed Error Message

Unknown error

Command Line

nextflow run '<https://github.com/nf-core/oncoanalyser>' -name distraught_mclean -params-file '<https://api.cloud.seqera.io/ephemeral/Iv8bfYxj-ebsgZjcG7isKA.json>' -with-tower -r dev -profile test_full

Aug 12, 2026 at 7:12 AM

Hi ! I have used the pipeline in wgts mode with donor samples and it works really well, thank you ! I tried to use it in targeted mode but it doesn't accept donor samples. Is there a restriction in the tools that accept donor samples in wgts (sage and amber if I understood well), and can't anymore in targeted mode ? In brief is it just not implemented in the pipeline (and could be in the future) or is it not implemented in the tools themselves ?

For generating custom panel-specific resources, what is the ideal scenario of tumor and normal samples to be included? The documentation suggests minimum 20 samples (tumor?). What happens if I have 10? Can these samples be part of the test data or do they have to be independent? And I’m also unclear about “Panel artefact PON”. Are these supposed to be tumor-specific artefacts from the above samples or germline events from the normal samples for filtering variants? Do I also need additional normal samples to create panel-specific germline events? If I also want to target RNA events in my panel, do I also need paired RNA-seq data for those >=20 samples for TPM normalisation? What if I use non-paired RNA samples? The doc says “files from a representative set of sample” but what if I have FFPE samples where RNA quality is bad and cannot be sequenced?

Hello everyone! I'm trying to run a

test_full

of oncoanalyser version 2.2.0 but

READ_ALIGNMENT_DNA:BWAMEM2_ALIGN

fails with no specific error (see error log in thread). So I was wondering, is this a known issue that was patched in the latest version? Couldn't find anything in the repository or documentation. It ran with

-profile test_full,singularity

and the following parameters:

mode: wgts
genome: GRCh38_hmf
create_stub_placeholders: false

Crossposting from an issue on Wigits/hmftools github in case anyone has solutions here: Hi all, I am having trouble processing one sample (WGS ~100x T, 40x N). Esvee raises an error "failed to process junction": Command output:

...
  16:11:33.339 [DEBUG] deleting 25 candidate BAMs
  16:11:33.520 [INFO ] 22482263 records written to BAM
  16:11:33.521 [DEBUG] writing sorted BAM: esvee/PANX_1341_Lv_M.esvee.prep.bam
  16:12:00.532 [DEBUG] writing sorted BAM: esvee/PANX_1341_Ly_R.esvee.prep.bam
  16:12:08.262 [INFO ] Esvee prep complete, mins(30.158)
  16:12:08.298 [DEBUG] processing tumor prep bam(s): [esvee/PANX_1341_Lv_M.esvee.prep.bam]
  16:12:08.298 [DEBUG] processing reference prep bam(s): [esvee/PANX_1341_Ly_R.esvee.prep.bam]
  16:12:08.344 [INFO ] writing to output directory(esvee/)
  16:12:08.365 [INFO ] fragment length bounds(min=151 max=779)
  16:12:11.166 [INFO ] loaded 461154 junctions, types(split=454597 discordant=6557 indel=13833 hotspot=3) from file: esvee/PANX_1341_Lv_M.esvee.prep.junction.tsv
  16:12:14.849 [DEBUG] primary assembly splits 269998 junction groups across 40 threads
  16:12:19.754 [WARN ] assembly(junc(chr1:2678382:-1) coords(extLen=37 refLen=134 refBasePos=2678515 len=171 juncIndex=37) support(38) mismatches(8)) excluding long remote region(chr1:2659634-2671618 reads(117) counts(mate=0 supp=0 disc=117) softClipMapQual(0))
  16:12:33.212 [DEBUG] processed 10000 junction groups, remaining(259998)
  16:12:37.415 [ERROR] failed to process junction(chr19:1161907:-1) with 249 reads

Command error:

...
  16:12:11.166 [INFO ] loaded 461154 junctions, types(split=454597 discordant=6557 indel=13833 hotspot=3) from file: esvee/PANX_1341_Lv_M.esvee.prep.junction.tsv
  16:12:14.849 [DEBUG] primary assembly splits 269998 junction groups across 40 threads
  16:12:19.754 [WARN ] assembly(junc(chr1:2678382:-1) coords(extLen=37 refLen=134 refBasePos=2678515 len=171 juncIndex=37) support(38) mismatches(8)) excluding long remote region(chr1:2659634-2671618 reads(117) counts(mate=0 supp=0 disc=117) softClipMapQual(0))
  16:12:33.212 [DEBUG] processed 10000 junction groups, remaining(259998)
  16:12:37.415 [ERROR] failed to process junction(chr19:1161907:-1) with 249 reads
  java.lang.ArrayIndexOutOfBoundsException: Index 61 out of bounds for length 61
  	at com.hartwig.hmftools.esvee.assembly.ExtReadParseState.currentBase(ExtReadParseState.java:76)
  	at com.hartwig.hmftools.esvee.assembly.ExtensionSeqBuilder.buildSequence(ExtensionSeqBuilder.java:263)
  	at com.hartwig.hmftools.esvee.assembly.ExtensionSeqBuilder.<init>(ExtensionSeqBuilder.java:131)
  	at com.hartwig.hmftools.esvee.assembly.JunctionAssembler.processJunction(JunctionAssembler.java:125)
  	at com.hartwig.hmftools.esvee.assembly.JunctionGroupAssembler.processJunctionGroup(JunctionGroupAssembler.java:186)
  	at com.hartwig.hmftools.esvee.assembly.JunctionGroupAssembler.run(JunctionGroupAssembler.java:113)

Any clues as to where to look at? Let me know if you need i.e. the esvee prep bam files or slices. This is ESVEE v1.2 on Oncoanalyser v2.3. I have run >100 samples in that cohort, this is the only one that fails. Thanks! Jose

✅ oncoanalyser/test_full completed successfully! 🎉 Run Name irreverent_meninsky Duration 5h 57m 2s Status ✅ Success Tasks Completed: 56, Failed: 1 Aug 29, 2026 at 1:12 PM

Hi <!here>, oncoanalyser 3.0.0 has been released, headline changes include: Sequencing data Support Ultima and Roche SBX / Axelios sequencing platforms New UMI processing via fastq-tools Predefined UMI configuration for selected panels Analysis Germline amplification calling ORANGE adjusted for driver reportability changes Improved noise detection and filtering, and purity fitting in tumor-only mode Single-step RNA alignment for each sample Read group overrides in the samplesheet New components MultiQC for aggregated QC reporting QSEE for detailed per-sample QC reporting with HMF cohort background SAGE visualiser for per-variant read support HTML reports Other Migrated to strict Nextflow syntax, workflow outputs, and channel topics Please see the oncoanalyser CHANGELOG.md and Release Notes for general details, and also the Hartwig WiGiTs_v3.0_ReleaseNotes.pdf Release Notes for the specific technical tooling changes The Hartwig team have also recently developed a detailed and helpful ORANGE User Manual for navigating, understanding, and interpreting results / findings. It's very well put together and worth reading for new and experienced users alike Thanks to everyone involved in this release and always special thanks to all the Hartwig folks for the immense amount of work done across their software suite and contributions to oncoanalyser

✅ oncoanalyser/test_full completed successfully! 🎉 Run Name agitated_visvesvaraya Duration 6h 11m 40s Status ✅ Success Tasks Completed: 56, Failed: 1 Sep 1, 2026 at 12:16 PM

Is there some details on where the files in the

hmf_resources

folder are, and where they come from?

Hi Simon, we have this starting document github.com/hartwigmedical/hmftools/blob/…/README_RESOURCES.md

happy to hear any feedback you may have otherwise

Thanks Rayan, that is a start. Is there any information as to how they were generated?

Yes, most of our resources are empirically generated using our own database and some potential additional manual curation by our medical team

they're ever-evolving, especially with user feedback

OncoKB™ Cancer Gene List has Gene Type as

ONCOGENE_AND_TSG

for some genes. However,

DriverGenePanel

for custom panel-specific resource accepts either

ONCO

or

TSG

. How to resolve this?

Hi @Ömer An, that's right, the

likelihoodType

field only accepts ONCO or TSG, and this choice changes how driver likelihood is scored. So for a genes listed as both in OncoKB you'll need to make a call yourself for the driver gene panel e.g. pick the role most relevant to your analysis, or fall back on the dominant role in biology / literature

I think

likelihoodType

picks a statistical model, not a biological label — ONCO expects clustered recurrent variants, TSG expects dispersed inactivation plus biallelic loss.

Hi! I have ran a case we have through both OA v2.2 and OA v3.0. In OA v2.2 the case had a QC PASS and a purity of 98% (13%-100%). In OA v3.0 The QC check fails with FAIL_NO_TUMOR and a purity of 100% (14%-100%). This also leads to the remaining of the Orange report being empty. • How can the two versions differ so significantly in the QC check? • Related to that, why is the purity range so large, and contradictory with the QC check? • Is there a way of forcing the Orange report generation?

I opened this PR to introduce bwa-mem3 to the pipeline. Users can add

--compat=bwa-mem2

to

ext.args

for BWAMEM3_ALIGN to get alignment records identical to bwa-mem2 v2.2.1 (the PG header line still names bwa-mem3, so output files are not byte-identical). I'm also happy to include a more explicit switch if the team prefers that strategy.

Hello oncoanalyser team, We have recently switched to the pipeline version 3.0.0 and wanted to ask if there is a way to publish the alignments as CRAM instead of BAM.

Hey oncoanalyser team! 👋 Thanks for such a wonderful tool! 🌟 I'm a bioinformatician part of Clinical Genomics Stockholm, and I've been looking into the SVs in the Purple VCF. We're aware that the SVs have a bit of an unconventional format required for Linx downstream, but to make it work for our downstream tools we've made some scripts to adapt the format, and it seems to be working quite well. The only thing left to handle is that the variants are represented as 2 different records for the same variant (which we're only used to seeing for BNDs), and we're thinking to add another script to merge the second record into the first, basically removing it and adding information to the INFO field of the first. But before proceeding with this, do you see any obvious issues with doing this what we should consider? 🤔

Hi! I was wondering if there is an easy way to provide a custom

known_fusion_data.38.csv

file to a run? Like a flag or similar?

Hi, would it be possible to include log2ratios in the cnv somatic table from purple? Made an issue here github.com/hartwigmedical/hmftools/issues/902 to track the request

Hi team! Thanks for your continued work on OncoAnalyser! With the 3.0.0 release, I was wondering whether the team's position on WES support has changed. I didn't see an explicit update on WES validation in the release notes. A couple of related questions: • For panel_resource_creation, does the ≥20-sample recommendation refer to patients, tumour/normal pairs, or total samples? • My understanding is that WES (like panel only run) would exclude features like TEAL, CHORD, CUPPA, SIGS, and possibly Orange reports. Is my understanding correct? I'm evaluating WES for larger-scale studies and trying to understand the technical feasibility. Any guidance on recommended resources, validation criteria, or existing WES users would be appreciated!

Hi team! We have observed that the mean ratio in cobalt.ratio.pcf files centered around 0 in OA2.2.0 (expected behaviour) while in version 3.0.0 this has changed and it now centers around 1. What’s the reason for this change?