Kondrashov’s Paper, Synergistic Epistasis, Soft Selection

Posted on by

Regarding Kondrashov’s paper Why have we not died 100 times over?, an internet “pupil” of mine who goes by the handle of “nomenmeum” asked what is synergistic epistasis and soft selection. The definition of synergistic epistasis seems to be the major issue as I don’t see the term much in literature. I don’t know exactly what it means. Does synergistic epistasis entail a change in S-coefficients?

For example in Brenda Andrews’ double and triple knockout experiments with yeast, the corresponding single knockout experiments had little-to-no noticeable effects, but several double and triple knockouts would clearly have deleterious effect when their component individual single knockouts did not in isolation.

Below is Brenda Andrews’ description of her experiment which I mentioned at Sandwalk in connection with the ENCODE 2015 planning meeting (Larry Moran knew Dr. Andrews as a graduate student at his school):

Brenda Andrews ENCODE 2015 Slides

So I would presume a double knockout scenario that was deleterious where the component individual knockout scenarios were neutral, the S-coefficients would change between the different scenarios. Is this an example of synergistic epistasis?

Kondrashov says:

“This paradox [i.e., humans existing for millions of years and yet not dying out 100 times over as a result of error catastrophe] cannot be resolved by invoking beneficial mutations or environmental fluctuations. Several possible resolutions are considered, including soft selection and synergistic epistasis among very slightly deleterious mutations.”

Thanks in advance for any responses.

NOTES:
There were two major ENCODE 2015 meetings, one for users, one for planners. I was in the users meeting, not in the planning meeting. Brenda Andrews presented at the planning meeting. Larry Moran knew Brenda Andrews and her work was discussed in passing here:
What did ENCODE Researchers Say on Reddit

8 thoughts on “Kondrashov’s Paper, Synergistic Epistasis, Soft Selection”

  1. I’m nomenmeum who asked Sal about synergistic epistasis. Thanks in advance for any responses.

  2. From Patrick Phillip’s 2008 review of epistasis:

    Much like the definitions of epistasis, there is a range of terms associated with particular forms of epistatic effects. Examples include synergistic, diminishing, antagonistic, aggravating, ameliorating, buffering, compensatory and reinforcing. Most of these refer to similar phenomena, which makes it difficult to understand what individual researchers mean when they use these terms. For example, synergistic epistasis occurs when an individual with a particular two-locus combination of alleles displays a phenotype beyond that expected from the individual effects of the alleles. If these are deleterious mutations then the phenotype is less than expected, but for positive mutations the phenotype is greater than expected. So sometimes synergistic epistasis means ‘extra good’ and sometimes it means ‘extra bad’. The field would benefit if all of these terms, which have context-dependent meanings, were replaced with two simple terms: positive epistasis and negative epistasis. Positive epistasis means that the phenotype is higher than expected and negative epistasis means that the phenotype is lower than expected. These two terms are preferable because their meaning is immediately clear and because it is the sign of the epistasis that matters in most evolutionary processes (such as the generation of linkage disequilibrium), not the change in relative direction of the effects of the individual loci.

  3. From Kondrashov 1994:

    Some data (MUM 1969; MALMBERG 1977) suggest that, rather than acting independently, deleterious alleles may act synergistically in the sense that each additional mutation causes a larger decline of relative fitness (see
    SHNOL and KONDRASHOV 1993).

    A few minutes searching on Google Scholar also results in a number of other papers that do a more-or-less reasonable job of defining the term.

  4. Dave,

    Thanks for the info. Would you feel comfortable saying the S-coefficients of the individual traits in the double knockout should be different than the S-coefficient in the component single knockouts. That’s how I would interpret it, but I felt not exactly comfortable stating that way without running the idea across evolutionary biologists like you.

    Thanks again.

  5. stcordova: Would you feel comfortable saying the S-coefficients of the individual traits in the double knockout should be different than the S-coefficient in the component single knockouts.

    The interesting thing here is that the strength of selection against any one deleterious allele depends on the frequency of the other. I guess there are several ways to model that, but most straightforward would be to introduce an additional interaction term that becomes non-zero when the deleterious alleles are combined.

  6. stcordova:
    Dave,

    Thanks for the info.Would you feel comfortable saying the S-coefficients of the individual traits in the double knockout should be different than the S-coefficient in the component single knockouts. That’s how I would interpret it, but I felt not exactly comfortable stating that way without running the idea across evolutionary biologists like you.

    I haven’t watched the video, but from your description that seems accurate enough. The idea is that, when epistasis is synergistic, the combined effect of each mutation (deleterious or harmful) when they are present at the same time is magnified beyond what would be expected if their effects were simply additive.

    There might be explanations other than synergistic epistasis for what is going on in the knockout experiments, but it seems likely that for at least one of the knockout alleles, its selection coefficient is probably going to be different when combined with the other allele than when it is present as only a single knockout.

  7. Dave Carlson:
    There might be explanations other than synergistic epistasis for what is going on in the knockout experiments, but it seems likely that for at least one of the knockout alleles, its selection coefficient is probably going to be different when combined with the other allele than when it is present as only a single knockout.

    The complication to watch out for is that the larger the total reduction in fitness is (reduction from having no knockouts), the more easily you will detect it statistically. Thus it would be tempting to conclude that there is no effect of a single knockout, when actually it is just that the effect was small enough that your statistical methods missed it.

    Also, you would want to actually estimate fitnesses, not just eyeball the organism and say “it seems to be perfectly normal”. The literature is full of such eyeball assessments, mostly by molecular and cell biologists. Population biologists know better than to do that.

  8. Dave:

    I haven’t watched the video, but from your description that seems accurate enough.

    Since the yeast genome is smallish, say a few thousand genes, a double knockout survey would involve a only a “few” million combinations. The hard part is the knockout.

    They put the combinations on slides and then simply measured the amount of light that could pass through the slide and then recorded it. For this to work well, the affect on fitness would have to be severe enough to be detected optically. The database was immense.

    The real aim was to see gene associations.

    Larry Moran commented that the ENCODE consortium would be thrilled to do this for human cell lines as it would guarantee research funding for decades. 🙂

Leave a Reply