As of now, failure in the assembler processes may trigger an error in the .status file but it does not trigger the error for the nextflow process. This leads to the assembly not being performed due to some reason and the resume cannot repeat the process, since it terminated successfully as far as nextflow knows.
The optional true tag should be removed from these processes to force the error in these processes and the error strategy could be changed to ignore, so that it does not compromise the rest of the pipeline. Coupled with the adaptive RAM directive in #26 , this should allow for a more robust and informative approach.