4. Settings - Organisms and simulation

Clicking on the settings button of the toolbar will launch a settings dialogue which has two tabs - one with settings for the organisms in the simulation, and one controlling the environment for the simulation.

4.1. Organisms

Number of characters/genome size

This defines the number of bits in character string for each organism, and ultimately the character number of the matrix output by the software.

Characters for species identification

This defines the number of characters used to identify species, between zero and the stated limit. By default this will update to be the same value as the genome size, but changing it to a smaller value allows a portion of characters that are not tied to species identity.

Characters for fitness calculation

This defines the number of characters used in the fitness algorithm, between zero and the stated limit. By default this will update to be the same value as the genome size, but changing it to a smaller value allows for a ‘non-coding’ portion of the genome, with a more drift-like/stochastic behaviour (although this portion will hitch-hike and thus could reflect patterns driven by the selection of organisms).

Rate of organism mutation

This is the rate of mutation for the organisms in the simulation, in units of mutations per hundred characters per iteration.

Stochastic layer

This tick box dictates whether a stochastic layer is used. The stochastic layer is a secondary string of bits that can be used to abstract the fitness calculation from the genome and its mutations, through a mechanism based on many-to-one mapping. When this box is ticked, on settings close, another dialogue is launched that allows you to - for every possible combination of 4 bits - define an output, i.e. 0 or 1, referred to here as the map. When the setting is enabled, each organism stores an internal genome four times the length of the organism’s gnome. During the simulation, mutations are applied to the stochastic layer (at the user defined rate per 100 characters), but all other operations employ the standard genome. When mutations are applied, the map is then used to convert bits from the stochastic layer into the genome that is used elsewhere (i.e. for every four stochastic bits, the output is 1 standard bit, based on the map). This abstraction process could be considered, for instance, a highly simplified parallel to e.g. ontogeny: it removes the fitness of the organism from direct control of its genome, and thus facilitates silent mutations and neutral theory of evolution type / less strongly adaptationist dynamics.

4.2. Simulations

Run mode radio

This has two options, run for iteration number or run to taxon number, which switches between the two options below.

Run to taxon number

If run mode is taxon number, a simulation will run until this number of species has evolved, and then terminate.

Run for iterations number

If run mode is iteration number, a simulation will run until this number of iterations, and then terminate.

Species difference

The hamming distance between a selected organism, post mutation, and either the character string when the species originated, or the last species to originate within the lineage (see paper for rationale) is used within the TREvoSim species concept. This setting defines the required hamming distance for a speciation to have occurred.

Unresolvable cut off

With low character numbers, especially when TREvoSim is set to strip out uninformative characters, terminals can have the same character string within a matrix, and thus be unresolvable. This setting defines upper limit for identical terminals. If there are more unresolvable taxa than this number in a single run, output files are not written and a warning is provided. If batch mode is underway, the current run is discarded and started again.

Fitness target

This is the target value for the count of 1s in the fitness calculation. See Garwood et al (2019) or Keating et al (2020) for details of the fitness algorithm. At low and high values (max being mask number multiplied by character number) fewer genomes will have peak fitness, at half max value, a larger number of peak fitness organisms exist. This can be quantified using fitness histogram menu command.

No selection

If this option is enabled, fitnesses are not calculated for digital organisms in the playing field, and instead every iteration and random individual is selected for replication.

Sansomian speciation

This option dictates when the character string of a species is recorded: at speciation, or at extinction. Sansomian speciation, the default, is to record the this when a species goes extinct. This ensures if it is a long-surviving species that the recorded characters string best reflects that closest to its sister group. When the option is not selected, genome is recorded at speciation.

Discard deleterious mutations

Optionally, TREvoSim can accept only mutations which are neutral, or improve the fitness of an organism.

Strip uninformative characters via subsampling

TREvoSim gives the option of writing matrices of only parsimony informative characters. When this option is checked, the software attempts to provide the requested character number of only informative characters. It achieves this by multiplying the number of characters and species difference by an empirically calculated factor (see below) at the start of a run. After a run has completed, informative characters are randomly subsampled to the requested number of characters; if there are not enough characters to achieve this, in batch mode the run is discarded and restarted, in single run mode an error message is provided.

Recalculate multiplication factor on dialogue close

The multiplication factor required to achieve a set number of informative characters via subsampling depends on the settings of any given run. Selecting this option calculates and sets this factor empirically after the settings dialogue is closed by conducting a ten-run batch and working out the proportion of informative characters within those runs. This needs to be recalculated after any settings are changed. If it is not, or has not been set, the software will run with a large factor, and thus be slower than necessary.