Zachary Kurtz

Can you try the `hugemb-patch` branch? It's hacky work-around but at seems to resolve the error your data. https://github.com/zdk123/SpiecEasi/tree/hugemb-patch

It looks like the issue is allocating a large enough vector in the underlying C++ code. `huge` attempts to save space by preallocating up to the theoretical maximum, but this...

This can happen when the "correct" lambda is not in the lambda path. So you should bump up `nlambda` to 50 or so (this will require 5x the number of...

Thanks... Will test macOS install on 4.2...

working on Big-Sur with an M1 chip. It looks like the wrong compiler is getting used though. I'll have to dig into this some more. Maybe use the pre-compiled package...

Thanks for the answers all. I should put this in the README.

Using glasso, you could convert the sparse inverse covariance matrix to a correlation matrix. An example of this in the README [using American Gut data ](https://github.com/zdk123/SpiecEasi#analysis-of-american-gut-data) ``` secor

you're passing in the `spiec.easi` object, not the re-fit adjacency matrix. Perhaps try `getRefit(se.gl.amgut2)`

maybe it doesn't like the sparse matrix output? Try `as.matrix`?

Thanks for the report. I did some googling and this "bad error message" may be a result of a too-long error being returned from an assertion. I do see from...