Section: New Results

Symbiont-host co-cladogenesis and co-evolution at the sequence and network levels

The problem here was to: (i) study the co-evolution of a set of hosts and their symbionts, and (ii) to understand the genetic architecture of a parasitic invasion by investigating the different phenotypes such invasion produces in the host.

Work on the first point took longer than initially planned but two papers are now submitted. In the first, titled “Co-phylogeny Reconstruction via an Approximate Bayesian Computation", we describe an algorithm (Coala ) for estimating the frequency of co-evolutionary events based on a likelihood-free approach. The benefits of this method are twofold: (1) it provides more confidence in the set of costs to be used in a reconciliation, and (2) it allows to estimate the frequency of the events in cases where the dataset consists of trees with a large number of taxa. We evaluate our method on simulated and on real datasets. We show that in both cases, for a same pair of host and parasite trees, different sets of frequencies for the events constitute equally probable solutions. Moreover, sometimes these sets lead to different parsimonious optimal reconciliations, in the sense of presenting a different number of the events. For this reason, it appears crucial to take this into account before attempting any further biological interpretation of such reconciliations. More generally, we also show that the set of frequencies can vary widely depending on the input host and parasite trees. Indiscriminately applying a standard vector of costs may thus not be a good strategy.

In the second submitted paper related to the study of co-evolution and titled “Eucalypt : Efficient tree reconciliation enumerator", we present a polynomial-delay algorithm for enumerating all optimal reconciliations. We show that in general many optimal solutions exist. We give an example where, for two pairs of host-parasite trees having each less than 40 leaves, the number of solutions is 2309, even when only time feasible solutions are kept. To facilitate their interpretation, those solutions are also classified in terms of the number of each event that they contain. This often enables to reduce considerably the number of different classes of solutions to examine further, but the number may remain high enough (16 for the same example). Depending on the cost vector, both numbers may increase considerably (for the same instance, to respectively 4080384 and 275).

Concerning the second question (genetic architecture of a parasitic invasion), one such phenotype is called “cytoplasmic incompatibility” (CI). Briefly, when a parasite invades a male host, it induces the death of the host's offspring unless the female is also infected. This has been explained by a toxin/antitoxin model that involves a toxin deposited by the parasites in the male's sperm inducing the death of the zygote unless neutralised by an antidote produced by the parasites in the egg. One toxin/antitoxin pair is usually linked to one genetic factor. Given a set of observed CIs, the question is how many genetic factors explain it. In its simplest form, this mathematically translates into, given a bipartite graph, finding its minimum biclique edge cover. One biclique corresponds to one factor. We had previously analysed the complexity of the problem and proposed an algorithm that was this year applied to a set of CI data from Culex pipiens [18] .