Hi all. A quick announcement that SLiM version 2.4.2 has just been released. This is a bug fix release for a serious bug; upgrading is STRONGLY RECOMMENDED FOR ALL USERS.
The bug causes mutation identifiers, which should be unique to each mutation across an entire model run, to instead be reused in a random fashion. For one thing, this means that the mutation ids given by SLiM's standard output commands may be duplicated and non-unique, causing downstream analysis issues. For another, because SLiM relies upon the mutation ids to identify segregating mutations within a sample, this bug means that the particular mutations that are output may themselves be incorrect; some mutations that are in fact segregating may not be included in the output.
These two problems mean that output from SLiM 2.4.1 cannot be trusted. This problem was also present in SLiM 2.4, but it was not present in SLiM 2.3 and earlier; it was introduced in 2.4. Model runs in version 2.3 and earlier are thus unaffected by this bug. Model runs done in 2.4 or 2.4.1 should be re-run.
My apologies for this bug. I test SLiM extensively, but it is very difficult to catch some types of bugs since SLiM is intrinsically stochastic in nature; there is no canonically correct output for a given SLiM model, making some types of testing problematic. As always when using third-party software, please do extensive sanity checks on the results you get from SLiM to make sure it is working as intended for you.
Thanks to Andrew Sackman for reporting the problem.