(Work in progress, j.l., 22.09.04)
Jacques Labonne (1), Bruno Parisi (1)
(1) UMR INRA - UPPA ECOBIOP Ecologie COmportementale et BIOlogie des Populations de poissons Station d'Hydrobiologie, Quartier Ibaron 64310 Saint-Pée-sur-Nivelle tel: +33 (0) 5 59 51 59 85
The Dynet module aims at modelling the effect of hydrographic networks dendritic structure on population demogenetics. The principle is to generate networks (using the RiverMaker modeltool) with various features, to assess these features, and then to simulate the behaviour of theoretical metapopulations in these networks.
The model integrates habitat spatial structure, river fragmentation, individual behaviour (movement, spawning, mortality). Many of the mechanisms are derived from the Bidasoa module. However, migration routines were dramatically simplified (stepping-stone model). The life cycle was corrected: there is no development delay, all individuals spawn. This simple biological model allows to study the impact of the network structure. Each individual has a genotype, a single neutral locus, used to track the genetics of populations during the simulations.
The network is build as in the Bidasoa module, but it is generated randomly using the RiverMaker modeltool. Each network may be saved in a file. At initialization, two files are required: a “network” file and a “species” file. The latter includes usual data about species and initial individuals. Some tools are available to quantify networks structure: see Rn and Rl indicators, but also the ReachConnectivity DataExtractor, that provides the distribution of neighbours numbers for each reach in a given range. The dynamcis is analysed with other DataExtractors such as FishPatchOccupancy (patch occupancy rate) or LocalExtinctionProbability (local extinction rate). Some DataExtractors for genetics are also available, like FishTimeFST (Wright FST measures), FishTimeFIS (intrapopluational fixation index), or the linear relationship between pairwise FST and geographical distances between patches.
The module utilizes frequently the the Genetics 2.0 library and supports compact mode for long simulations. To retain the pertinent data alont time in compact mode, the module uses a “Recorder” class, here renamed “DynRecorder”. This class is merely a data structure where various informations and matrices can be recorded. Only data of interest are written, there is no access to the different steps (due to compact mode), but a lot of memory is still spared.
This module is presently stand-alone, not yet free to distribute, and only necessitates the CAPSIS4 software (which is licenced under GNU Lesser GPL).
Inventory file description:
The Guppy module uses a text input format that includes habitat data, one line per reach, fragmentation data, one line per reach, and populational data, one line per fish.
Value : species number Name : species name GeneticMap : genetic map to provide if not deduced from individual data AllelesNuclear : available nuclear alleles AllelesMCytoplasmic : available maternal cytoplasmic alleles AllelesPCytoplasmic : available paternal cytoplasmic alleles (here used as Y-linked alleles)
ID : unique ID address : string that identifies each reach in the connexion structure of the network order : Strahler order length : length (m) meanWidth : mean width (m) fish : IDs of the fish intially contained in the reach Vertices : vertex3D collection to draw the river
ID : unique ID passability : passability (from 0 = no obstacle to 3 = full blocking) origin : vertex3D for location address : string that identifies each weir in the connexion structure of the network
ID : unique ID sex : sex age : age (the time steps is twelve days) nucDNA : classical nuclear genotype mCytDNA : mitochondrial genotype pCytDNA : Y-linked genotype
mId: mother ID pId : father ID creationDate : birth date
To analyze population dynamics of the Guppy module, one can use Visus and graphics from CAPSIS. A specific output format for script simulation results might be in our wish list.