Error: all penalty factors are <= 0

[lbeltrame]

I'm running siglasso with some samples that may be problematic, and I get errors like

Error: from glmnet Fortran code (error code 10000); All penalty factors are <= 0

I'm wondering if siglasso should just skip these (and fill the relevant sample data with NA). At the same time, is there any way to find why these samples are problematic? A few of these had too few variants (and were promptly removed), but others have reasonable numbers (50+).