Hola again @ajfaure,
Thanks again for checking the recent cutadapt version upgrade (#23)!
Now another challenge has come up during our efforts to build DiMSum containers: ARM64 processor architectures like e.g. the Mac M1/M2/M3 series require substantially newer R versions than 4.0.0. We are able to compile DiMSum singularity images with R v4.4.3, but have started running into what looks like a general R environment incompatibility on our small scale test data (which works perfectly fine with the default DiMSum + R v4.0.0).
Note that the same error had already been observed by @roserao some time ago (#16):
******* DiMSum STAGE 4 (STEAM): PROCESS VARIANT SEQUENCES *******
Loading variant count files:
counts_merged.tsv
Processing...
counts_merged.tsv
Processing merged variants...
Done
Saving merged variant data...
Done
There were problems while running 'dimsum__merge_report'
There were problems while running 'dimsum__diagnostics_report'
Error in dimsum_stage_merge(dimsum_meta = pipeline[["3_tally"]], merge_outpath = pipeline[["3_tally"]][["project_path"]], :
object 'dimsum_meta_new_report' not found
Calls: dimsum -> dimsum_stage_merge
Execution halted
Do you have a feel for what it could be that is triggering this? I'd be happy to get involved with any debugging efforts, though am finding it a bit challenging to identify the right starting point and main underlying source.
Note that our test is only from stage 4 onwards, hence not sure if there may be other potential R >4.0.0 warnings upstream. Sharing our toy data files and run details here:
counts_merged.tsv
experimentalDesign.tsv
2025-09-30_run_DiMSum_test.sh
@BenjaminWehnert1008 @tlmighell @wtsi-hgi @ewels
Hola again @ajfaure,
Thanks again for checking the recent cutadapt version upgrade (#23)!
Now another challenge has come up during our efforts to build DiMSum containers: ARM64 processor architectures like e.g. the Mac M1/M2/M3 series require substantially newer R versions than 4.0.0. We are able to compile DiMSum singularity images with R v4.4.3, but have started running into what looks like a general R environment incompatibility on our small scale test data (which works perfectly fine with the default DiMSum + R v4.0.0).
Note that the same error had already been observed by @roserao some time ago (#16):
Do you have a feel for what it could be that is triggering this? I'd be happy to get involved with any debugging efforts, though am finding it a bit challenging to identify the right starting point and main underlying source.
Note that our test is only from stage 4 onwards, hence not sure if there may be other potential R >4.0.0 warnings upstream. Sharing our toy data files and run details here:
counts_merged.tsv
experimentalDesign.tsv
2025-09-30_run_DiMSum_test.sh
@BenjaminWehnert1008 @tlmighell @wtsi-hgi @ewels