This new review explains why soft sweeps are the bane — and the baseline — of ecological genetics

(Flickr: andrew)

If you’ve done ecological genetics research in the last decade, you’ve almost certainly cited a series of papers by Pleuni Pennings and Joachim Hermisson, which broke down the problem of soft selective sweeps. Pennings and Hermisson have revisited soft sweeps in a big, detailed new review article for Methods in Ecology and Evolution, which pulls together more than a decade of research following the original studies, and makes a good case that everyone’s favorite excuse for a less-than-dramatic genome scan result is not going away any time soon.

First, what exactly is a soft sweep? Well, it’s a selective sweep that is … not hard. The original papers addressed a couple different ways that natural selection might fail to produce the classic signature of a “hard” selective sweep — in which a single advantageous genetic variant spreads through a population over a few generations, eventually becoming the only variant present — but didn’t quite line them up for comparison. In the new review, Hermisson and Pennings do this very explicitly.

First, they clearly define a hard sweep as a case where a population encounters a new selective environment, and a new beneficial mutation appears, then sweeps to fixation. By the end of the hard sweep, everyone in the population has a copy of the beneficial variant descended from that first mutation — the time to the common ancestor of the “fixed” variant is less than the time since the population started experiencing selection to favor that variant. This means that the region of genetic sequence linked to that variant is much less diverse than the rest of the genome, which makes it easy to pick out if you sample and sequence the population.

In contrast, soft sweeps happen under conditions when beneficial variants become more common in the population under conditions that don’t reduce nearby sequence diversity as strongly as an equivalent hard sweep would. The first way this can happen is if the beneficial variant already exists in the population when selection begins to favor it — the sweep is fueled by standing genetic variation. If you know a little coalescent theory, this is intuitive: that longer time to common ancestry means more time for acquisition of mutations in the region linked to the beneficial variant, and more time for the variant to recombine onto new backgrounds. (If this isn’t intuitive … well, we’re overdue for a primer on the coalescent, now that I think about it. Hold that thought.)

A single-origin soft sweep occurs when the common ancestor of a variant that has swept to fixation (TMRCA) is older than the period during which selection has favored that variant (TS). (Detail of Figure 1, Hermisson and Pennings 2017)

This isn’t the only way for a sweep to soften, though. Even if mutation doesn’t create the beneficial variant before selection begins to act, it might be that mutation creates multiple beneficial variants. That, too, can result in multiple genetic backgrounds — a different one for each origin of the beneficial variant — increasing in frequency, and recombining with each other, until the variant has fixed. Multiple-origin soft sweeps have their own properties that are distinct from single-origin soft sweeps, though both leave “footprints” in the diversity of populations that are harder to detect and understand than a nice, simple hard sweep.

Hermisson and Pennings spend the bulk of the review drawing on past population genetic theory to delineate conditions under which each kind of selective sweep is likely. Hard sweeps are more likely when the effective population size prior to selection is smaller, when mutation is more rare; and in cases when selection is actually against a variant before new environmental conditions make it advantageous. Soft sweeps are more likely when any of those conditions aren’t met, with multiple-origin soft sweeps expected under a rather wider range of specific parameter values than single-origin ones. This is a paper where it pays to follow the math closely, but key results are nicely spelled out in the text and illustrated. Pennings, who has done a lot of neat video communication of her scientific work, has also produced a video to accompany the paper, which works out some of the key conclusions in a sort of chalkboard animation.

Hermisson and Pennings make a credible case that conditions for soft sweeps are, indeed, pretty common. They consider recent work in three major empirical systems — Drosophila, humans, and various microbes — all of which have well documented cases of soft sweeps. In Drosophila, which has sufficiently large populations that the theory outlined by the authors predicts a mixture of hard sweeps and multiple-origin soft sweeps, soft sweeps have been documented in multiple parts of the genome, including in the context of a haplotype-frequency-based test for selective sweeps that can cope with (some) softness. Humans have, historically, had pretty small effective population sizes, but we have also colonized and created a lot of new environments in rapid succession, so that when adaptation occurs it is more likely fueled by standing variation. Microbial cases include HIV adapting to antiviral drugs within single human hosts — some of Penning’s own empirical work has been with HIV.

The full review is well worth your time — beyond these highlights, Hermisson and Pennings dig into cases in which hard and soft sweeps might leave ambiguous genomic footprints, and the possibility of geographic soft sweeps, which should be of particular interest to molecular ecologists. A great deal of the trouble with soft sweeps boils down to the fundamental puzzle of evolutionary biology, reconstructing what happened in the past from the variation we see around us today. This review pulls together a lot of interesting work toward solving one particularly tricky part of that puzzle.


Garud NR, PW Messer, EO Buzbas, and DA Petrov, 2015. Recent selective sweeps in North American Drosophila melanogaster show signatures of soft sweeps. PLOS Genetics 11(2): e1005004. doi: 10.1371/journal.pgen.1005004

Hermisson J. and PS Pennings. 2005. Soft sweeps. Genetics 169(4): 2335-2352. doi: 10.1534/genetics.104.036947

Hermisson J. and PS Pennings. 2017. Soft sweeps and beyond: Understanding the patterns and probabilities of selection footprints under rapid adaptation. Methods in Ecology and Evolution 8(6): 700-716. doi: 10.1111/2041-210X.12808

Pennings, PS. and J Hermisson. 2006. Soft sweeps II — Molecular population genetics of adaptation from recurrent mutation or migration. Molecular Biology and Evolution 23(5):1076-1084. doi: 10.1093/molbev/msj117

Pennings PS. and J Hermisson. 2006. Soft sweeps III: The signature of positive selection from recurrent mutation. PLOS Genetics 2(12): e186. doi: 10.1371/journal.pgen.0020186


About Jeremy Yoder

Jeremy Yoder is a postdoctoral associate in the Department of Forest and Conservation Sciences at the University of British Columbia. He also blogs at Denim and Tweed, and tweets under the handle @jbyoder.
This entry was posted in adaptation, evolution, genomics and tagged , , . Bookmark the permalink.