SLiM 3.4 released, including new QtSLiM graphical modeling environment for Linux

99 views
Skip to first unread message

Ben Haller

unread,
May 12, 2020, 9:58:36 AM5/12/20
to slim-announce
Hi everyone.  We have just released SLiM 3.4.  This is a major release.

The big exciting news, and the reason this is a major release, is that SLiM now has a graphical modeling environment for Linux: QtSLiM.  QtSLiM is a port of SLiMgui from the Mac-only "Cocoa" object kit to a cross-platform object kit called "Qt" (pronounced "cute", so QtSLiM is "cute slim").  QtSLiM has most of SLiMgui's features, and is ready to use now.  Instructions for building and installing it on Linux are in section 2.4 of the new manual.  We have tested it on Ubuntu 18.04 LTS and 20.04 LTS, as well as on Debian and Arch, and it seems to run well on all.  A few common difficulties are discussed in section 2.4.3 of the manual.

QtSLiM also runs on macOS, since Qt is cross-platform.  In fact, we plan that it will become the new graphical modeling environment for all SLiM users, at which time it will adopt the name "SLiMgui" and the old SLiMgui will become obsolete.  We have not yet made that transition, however, because we would like QtSLiM to receive some testing from all of you first.  Mac users who want to try QtSLiM can download it from http://benhaller.com/slim/QtSLiM.app.zip; just unzip the downloaded file and drag QtSLiM.app to your /Applications folder.  If anything, it seems to be a bit *more* stable than SLiMgui on macOS 10.15, so we hope lots of Mac users will try it and give us feedback.

Besides QtSLiM, changes in SLiM 3.4 are fairly minor:

- writeFile() and writeTempFile() can now write .gz-compressed data directly to disk with compress=T

- qnorm(), dbeta(), dexp(), and dgamma() functions have been added to Eidos (quantile and density functions for various distributions)

- you can now optionally omit the generating script from SLiM's .trees provenance entries, which is nice if your scripts are very large; instead you can use a new SHA-256 hash to identify the generating script

- a new recipe, 16.18, shows how to build a spatial epidemiological S-I-R model in SLiM (unrelated to COVID-19; this has been on my to-do list for more than a year)

There are also a few bug fixes:

- recipes 9.5.2 and 9.5.3 have been modified to fix a bug involving fitness calculations in selective sweep models; if you might have derived your own model code from those recipes, you should have a look at the new recipes and discussion in the manual

- other bugs that you would know if you had hit (a crash, two fatal errors, a display glitch in SLiMgui)

SLiM 3.4 is available for installation under conda-forge; see section 2.5 of the manual.

This upgrade is recommended for all users, since it fixes a few serious bugs and should be quite stable.  This release preserves backward reproducibility (i.e., the same model, with the same seed, will produce the same results as SLiM 3.3.2, apart from the bugs fixed) and backward compatibility (models that ran under SLiM 3.3.2 should not require changes to their code).

You can obtain SLiM 3.4 from the SLiM home page at http://messerlab.org/slim/; note that the manuals, recipes, and reference sheets have also undergone revisions.

If you're a beginner in SLiM, you might want to check out our recent paper "Evolutionary modeling in SLiM 3 for beginners" (http://dx.doi.org/10.1093/molbev/msy237).

If you have any questions, comments, etc., please use the slim-discuss group for that.  Thanks, and happy modeling!
Cheers,

Benjamin C. Haller
Messer Lab
Cornell University

Reply all
Reply to author
Forward
0 new messages