Escape From Evolutionary Stasis by Transposon-mediated Deleterious Mutations

J. theor. Biol. (1997) 186, 441–447

Escape From Evolutionary Stasis by Transposon-mediated Deleterious Mutations J MF*‡  G K† *School of Biological Sciences, University of Surrey, Guildford, Surrey, GU2 5XH; and †Weeks Computing Services, 6 Langley St., London WC2H 9JA, U.K. (Received on 13 September 1996, Accepted in revised form on 14 January 1997)

Evolution within a rugged fitness landscape is limited by the tendency for organisms to become trapped on local optima resulting in evolutionary stasis. It is presently unclear how founder populations escape from an adaptive peak to found a new species. Insertion sequences, transposons and other mobile DNA elements are found in all species of eukaryotes, bacteria and archaebacteria, where they have been sought and are usually considered to be genomic parasites or selfish genes. However, many transposons and other mobile repetitive DNA are remarkably species or phyla-specific, indicating that infection with transposable elements coincides with speciation events and is involved in promoting evolutionary change. We propose here a model in which transposable elements are involved in speciation events by their ability to produce irreversible deleterious mutations that promote escape from evolutionary stasis. We have constructed a genetic algorithm designed to model both spontaneous and transposon-mediated mutations in populations of asexual digital organisms. We use this model to investigate the effect of transposon-mediated mutations on the rate of evolution of digital organisms as they compete for resources within an artificial adaptive landscape. In the absence of transposon mutations the seed organisms quickly evolve to occupy the nearest adaptive peak but thereafter evolutionary stasis ensues and adjacent empty peaks are left unoccupied. In the presence of transposon mutations, evolution is again dominated by stasis but is punctuated by bursts of rapid evolution in which consecutive unoccupied adaptive peaks are filled with organisms derived from single transposition events. Rapid evolutionary events leading to founding of new biological species, may be similarly initiated by irreversible deleterious mutations induced by transposition. 7 1997 Academic Press Limited

Introduction Wright (1932) introduced the concept of an adaptive landscape in which each point on the surface of the landscape corresponds to a particular genotype with an associated fitness and neighbouring points differ by single mutations. Evolution is envisaged as a walk across the adaptive landscape towards peaks of higher fitness. However, the ability of organisms to reach global fitness maxima in a rugged, many-peaked adaptive or fitness landscape is severely restricted by the tendency for any random walk to climb towards a local optimum and thereafter become trapped ‡Author to whom correspondence should be addressed 0022–5193/97/120441 + 07 $25.00/0/jt970403

(Eigen, 1985; Kauffman, 1993; Kauffman & Johnsen, 1991). In this situation virtually all single mutations will result in descendants that are less fit than the parental genotype. Since most single mutations may be reversed by back-mutations, compensatory mutations and/or sexual recombination (Chao, 1991; Stephan et al., 1993), there will be a tendency for subsequent generations of mutants to reverse the initial phenotypic shift, resulting in (phenotypic) evolutionary stasis. Under natural selective conditions, members of a biological species are therefore likely to remain tightly clustered around the local optimum for the species. The origin of new species must involve means for crossing the valleys of a fitness landscape. Simpson (1944) suggested that most higher 7 1997 Academic Press Limited

442

.   . 

phyla arise through rapid crossing of adaptive thresholds, passing from one biological niche to another, a phenomenon he termed macroevolution. More recently, Gould & Eldridge (1977, 1993), examining morphological change in the fossil record, claimed that species change very little during their existence but that this stasis is punctuated by rapid evolutionary change giving rise to new species. Comparison of sequences of bacterial ribosomal RNA genes has led Woose (Woose et al., 1985; Woose, 1987), to suggest that ‘‘major bacterial groups come into being through episodic rapid evolution’’. Macroevolution, remains controversial, however, and proposed genetic mechanisms that could produce rapid evolution are largely speculative. Insertion sequences, transposons and retrotransposons are ubiquitous mobile genetic elements (henceforth referred to generically as transposons) found in bacteria, archae eukaryotes. Although many transposons carry additional genes that provide a selective advantage to the host (e.g. antibiotic-resistance), most elements appear to be cryptic and are usually considered to be genomic parasites or selfish genes (Dawkins, 1976; Doolittle & Sapienza, 1980; Orgel & Crick, 1980). However, many repetitive DNA sequences are remarkably species or genera-specific: e.g., IS200 in Salmonella (Lam & Roth, 1983), IS492 in Pseudomonas atlantica, IS900 in Mycobacterium paratuberculosis (Green et al., 1989), IS6110 in Mycobacterium tuberculosis, IS481 in Bordetella pertussis (DeShazer et al., 1994), the alu repeat family in primate (Britten, 1994), the LINE family of retrotransposons in mammals (Smit et al., 1995), the Tc1 transposon in Caenorhabditis elegans (Abad et al., 1991), the Afa family of tandem repeats in Triticeae (Nagaki et al., 1995). A recent analysis of insertions of repeated elements into eukaryote genomes found evidence for specific insertions affecting gene regulation in the human, mouse and sea urchin genomes (Britten, 1996). This pattern of transposon-(or repetitive element) signatures for phyla suggests that infection with the transposon occurred at a level very close to the origin of the phylum and may have been involved in promoting the evolutionary change that led to speciation. Rose & Doolittle (1983) suggested that rapid speciation events may involve hybrid sterility mediated by ‘‘genomic disease’’ caused by infection of a genome by a transposon. This mechanism is thought to be involved in hybrid dysgenesis in Drosophila (Bingham et al., 1982), but has not been described elsewhere and is not relevant to evolution in the essentially asexual prokaryotes. The potential of transposons to insert at multiple sites and simultaneously modify multiple

loci, thereby causing major structural and regulatory changes in a single event, has also been proposed as a mechanism underlying rapid evolutionary events (Chao et al., 1983; Syvanen, 1984; Temin, 1982). However, this scenario is essentially a variation on the hopeful monster theory whereby new species are proposed to arise ‘‘full-blown as bizarre organisms quite distinct from their progenitors’’ (Goldschmidt, 1940). Such macro-mutational theories are now considered highly improbable due to the very low probability of large evolutionary jumps being anything other than severely deleterious and probably lethal (Dawkins, 1988; Kauffman, 1993). We propose here an alternative model whereby transposons are proposed to initiate speciation events by virtue of their ability to cause irreversible deleterious mutations that promote escape from evolutionary stasis. We demonstrate the feasibility of this model using a genetic algorithm to model evolution of asexual organisms within an artificial ecosystem (Conrad & Pattee, 1970). Model and Results We have constructed a genetic algorith, CREATURES, in which a string of characters, represent the genome of an asexual digital organism or ‘‘creature’’, e.g. BBBBB represents the genome of a hypothetical creature with five genes, each gene having 26 possible alleles (A–Z). Gene alleles that are alphabetically close (e.g., D & E) are considered to be homologous in function and phylogenetically related. The genome of the creatures are allowed to evolve by both micro-mutations (analogous to base-substitutions, small deletions, insertions and rearrangements, expansion and contraction of triplet repeats, etc.) and transposon-mutations. Micro-mutations cause a single gene character to shift by a single increment (e.g. BBBBB : BBBCB) at a defined rate—the micro-mutation rate. Insertion of a transposable element into a host chromosome will usually result in inactivation of the gene it inserts into (insertional inactivation). This is the predominant genetic effect of transposon insertion in bacteria (Chao et al., 1983; Kleckner, 1990) and our algorithm models this aspect of transposon behaviour. After transposon insertion, the inactivated gene-transposon locus will (from the host’s viewpoint) effectively be random genetic information that will be free to evolve in new directions. We have therefore modelled infection with a transposable element in our creature’s genome, by allowing transposon mutations in which single genes shift by random increments, (e.g., BBBBB : BTBBB) at a defined rate—the transposon mutation

    rate. A register recording both micro-mutation and transposon mutation rates are attached to each creature (essentially behaving as two extra genes that determine the mutation rates). The rates may be fixed for each creature or may be allowed to vary. This variation will, like the creature’s other genes, be subject to selection. We have modelled a multidimensional adaptive landscape consisting of peaks where finite quantities of ‘‘food‘‘ are located. The coordinates of the landscape correspond to the creature’s genome, i.e., the creature KKKKK will occupy the pinnacle of the peak KKKKK. Distance is a measure of the numerical genetic distance between a creature’s genome and the peak’s coordinates—how many micro-mutations would be required to ‘‘walk’’ that distance (e.g., the distance between BBBBB and CCCBC is four units). However, since genes do not function in isolation but are subject to multiple epistatic interactions with other genes, in our model, genetic distance is measured within an ‘‘epistatic window’’ of adjacent genes that are considered to form a functionally-related group. Mutations in one gene affects the fitness of all of the epistatically-related genes. Each peak is defined by: peak coordinates, food, regeneration (a measure of how fast the peak’s food is replenished after being consumed by the creatures) and pressure P. The proximity P, is a measure of the distance, raised to the power of the pressure. Proximity of a creature to a peak determines the chance it has of obtaining food from the peak. Therefore, pressure is a measure of the gradient of

443

abundance of food surrounding the peak’s coordinates—the gradient of the peak. Figure 1(a) shows the shape of a two-dimensional adaptive landscape (two character creatures) with five peaks, each with pressures of 50 and Fig. 1(b) shows the effect of the pressure of the peaks to 5. A world is seeded with a finite number of creatures decreasing of defined genome (e.g., 50 × AAAAA). Creatures compete for food and thereby replication success. At each round of replication the creature’s genomes will be subject to mutations according to the mutation rates.     A ‘‘world’’ run was usually initiated with a seed population of about 50 5–10 gene creatures with genomes set at a e.g., AAAAA (the behaviour of ten-character creatures was essentially the same but runs took far longer) and micro-mutation rate set to an initial rate of between 10 − 3–10 − 1. The transposon mutation rate was set to zero. A ‘‘capture peak’’ (pressure usually set to 20) was placed adjacent to the seed (e.g., BBBBB) with food set at 200 (supports maximum population of 200 creatures) to capture and maintain the descendants of the seed creatures (if the capture peak was positioned too close to the seed then its food was completely consumed—it suffered extinction). Eight additional peaks with randomlygenerated coordinates and randomly-allocated pressure values (5–50) were generated to construct a rugged five-dimensional fitness landscape. The ability of the creatures to evolve and fill the available empty peaks was monitored up to 50000 generations. The

F. 1. Two-dimensional adaptive landscape constructed using program CREATURES showing the proximity at each grid-point with five two-character peaks and pressures set to (a) 50 and (b) 5.

.   . 

444 (a) 700

Population

600 500 400 300 200 100 0

1

20

2000 12000 22000 31000 41000 Generation

(b) Rate of evolution

0.30 0.25 0.20 0.15 0.10 0.05 0

1

20

2000 12000 22000 31000 41000

Micro-mutation rate

Generation

(c) 0.006 0.005 0.004 0.003 0.002 0.001 0

1

20

2000 12000 22000 31000 41000 Generation

F. 2. Parallel runs of CREATURES with transposon mutation rate set to zero (broken line), and transposon mutation rate set to 10 − 3 per gene per generation (filed line), showing (a) world population, (b) rate of evolution and (c) micromutation rate, all plotted against the number of generations.

simplest measure of the evolutionary success of the creatures was the eventual population size of the world which was a measure of how many peaks were filled. Seed creatures would rapidly evolve towards the capture peak, obtaining food and replicating to fill the peak. Thereafter the population on the capture peak would remain stable, fluctuating at around two-thirds of the maximum population for the peak [Fig. 2(a)]. If the creature’s micro-mutation rates was allowed to vary from an initial rate of 0.005, it dropped rapidly to very low values (Q10 − 6) [Fig. 2(c)]. This was presumably a result of those creatures with the lowest mutation rates being most

successful in maintaining themselves on the peak against evolutionary drift (Stephan et al., 1993). Thereafter the creatures occupied a narrow band close to the pinnacle of the capture peak and the subsequent evolution of the world was dominated by stasis—low and static evolution rate [Fig. 2(b)], no invasion of empty peaks and static population size. Fixing the micro-mutation rate, even at relatively high values (e.g., 0.05) did not significantly increase the rate of evolution and creatures remained trapped on peaks.    Worlds were initiated as above but containing creatures with the transposon mutation rate q0 for all creatures. If the transposon mutation rate was allowed to drift, then it would rapidly be driven to very low rates, in exactly the same way as the micro-mutation rate, as discussed previously. Thereafter, the world would be dominated by stasis—very low evolution rates and no filling of empty peaks. However, although in biological organisms the micro-mutation rate may be minimised by selection for high fidelity of DNA polymerase enzymes and efficient systems for recognising and repairing DNA base mis-matches; the transposon mutation rate is not under the same endogenous control. Transposons are exogenous elements that may be delivered to bacterial host genomes by virus infection, conjugation, or DNA transformation. The rate at which these events occur is not necessarily under the control of the host and therefore not subject to host-level selection— transposons behave as ‘‘selfish genes’’. We have set the transposon mutation rate to a constant value, to mimic this relatively constant acquisition of exogenous transposable elements. Note that fixing the micro-mutation rate did not significantly change the behaviour of the worlds. With a fixed transposon mutation rate, creatures typically escaped evolutionary stasis (Fig. 2). Invasion of empty peaks was initiated by a transposon mutation resulting in a loss of fitness. Thereafter, the debilitated mutant (if it survived) rapidly evolved to find and fill a series of vacant adaptive peaks. Populations inhabiting newly-filled peaks were invariably clonal—they were derived from a single transposition event (recorded as the biggest mutational jump for the closest genome). A series of worlds were run to evaluate the overall behaviour of the system, varying the transposon mutation rate with the micro-mutation rate also allowed to vary (initialised at 0.02) or held constant (at 0.02). A series of 40 worlds each with a new set of random peaks were run and the final population

    sizes attained after 50000 generations are presented in Fig. 3. All eight runs with transposons mutation rate set to zero were characterised by evolutionary stasis with only the capture peak being filled with creatures. However, escape from evolutionary stasis was readily achieved in most worlds with transposon mutations. Escape from stasis was detected in some worlds with the transposon mutation rate set to 0.0005, a value 40 times lower than the micro-mutation rate in those runs. Worlds were also run with mixed populations of transposon-carrying and transposon-free creatures. In these, vacant peaks were always filled by the transposon-carrying creatures (data not shown). However, it should be noted that the model does not allow infection of transposon-free creatures with transposons.

Discussion The genetic algorithm described here models many aspects of the evolutionary behaviour of biological systems. Although we believe that the evolutionary potential of transposons will be similar in most biological systems, replication of our digital creatures is asexual (although we assume the existence of parasexual processes involved in delivery of transposons) and does not involve recombination. The model 600

Final population

500

400

300

200

100

0

0.0005

0.002

0.005

0.02

Transposon mutation rate F. 3. Final (averaged over last 10 generations) population sizes of series of CREATURES worlds run with various transposon mutations rates. Eight world runs were performed for each transposon mutation rate point. Circles indicate runs with fixed micromutation rate (10 − 2) and squares indicate worlds where the micro-mutation rate was initialised at 10 − 2 but thereafter allowed to vary.

445

is therefore only directly applicable to asexual organisms. It should be remembered however that asexual replication is a widespread form of procreation found in prokaryotes and many eukaryotes including arthropods, plants and protozoa. Selective pressure maintains both our digital creatures and biological organisms on narrow adaptive peaks. Micro-mutations (base-substitutions, triplet expansion/contraction, small deletions, insertions, rearrangements, etc.) are capable of modifying the genotype and phenotype of creatures but since most organisms are highly adapted for the ecological niche they inhabit, mutants will invariably be less fit than their parents. The reversibility of most micro-mutations ensures that selective pressure tends to reverse any movement away from the peak and prevent escape from the peak. This ‘‘evolutionary drag’’ of the peak will be reinforced by powerful selection towards low mutation rates on adaptive peaks. The inevitable outcome is that, in the absence of transposon mutations, creatures become trapped on adaptive peaks and adjacent ecological niches (adaptive peaks) will remain vacant. In contrast to micro-mutations, insertion of a transposable element into the chromosome will usually be irreversible—rates of excision are extremely low (and may leave a small target-site duplication ‘‘remnant’’ that continues to disrupt the gene) and insertion generally involves multiple loci. Although most transposon mutations will be deleterious or lethal, some mutants will occasionally be debilitated but manage to survive and replicate. These debilitated mutants will have effectively escaped the ‘‘evolutionary drag’’ of the parental peak. Many micro-mutations will now be adaptive—the mutant offspring may be more fit than their parents. Selective pressure on the micro-mutation rate will be relaxed, allowing it to rise and facilitate the mutant lines, ability to search the adaptive landscape, invade vacant ecological niches (adaptive peaks) and speciate. This model is illustrated in Fig. 4. It is important to emphasise that neither in our model nor in real biological systems is there any likelihood that transposon mutations will be capable of finding new peaks in a single jump—they are not a class of macromutations. The transposon-derived mutant effectively jumps randomly into nearly empty sequence space. It is only the subsequent ‘‘micro-evolution’’, that is capable of moulding both our creatures and biological creatures into new species or phyla. As demonstrated in our model, the descendants of the original debilitated mutant will carry a clone, species or phyla-specific transposon signature. Our model therefore predicts the presence of transposons

.   . 

446

F. 4. Model for escape from evolutionary stasis by transposon-mediated deleterious mutations.

(or repetitive element) signatures in extanct creatures and suggests that they are the relics of bursts of rapid evolution through transposon-mediated debilitated mutants. Our model predicts that evolutionary innovations are likely to be preceded by debilitating mutations. It is therefore notable that many extant phyla are characterised by loss of a phenotypic character, e.g., loss of cell wall in the mycoplasmas, loss of photosynthetic ability in many members of the proteobacteria (Woose, 1987). A further characteristic of our model is that evolution is dominated by periods of stasis separated by bursts of rapid evolution or macroevolution (Simpson, 1944). Transposons may therefore play an important role in evolution by promoting escape from evolutionary stasis.

REFERENCES A, P., Q, C., T, S., P, C., C S, P., A, M. et al. (1991). Sequences homologous to Tc(s) transposable elements of Caenorhabditis elegans are widely distributed in the phylum nematoda. J. Mol. Evol. 33, 251–258. B, P. M., K, M. G. & R, G. M. (1982). The molecular basis for P-M hybrid dysgenesis: the role of the P element, a P strain-specific transposon family. Cell 29, 995–1004. B, R. J. (1994). Evidence that most human Alu sequences were inserted in a process that ceased about 30 million years ago. Proc. Natl Acad. Sci. U.S.A. 91, 6148–6150. B, R. J. (1996). DNA sequence insertion and evolutionary variation in gene regulation. Proc. Natl Acad. Sci. U.S.A. 93, 9374–9377. C, L., V, C., S, B. B. & C, E. C. (1983). Transposable elements as mutator genes in evolution. Nature 303, 633–635. C, L. (1991). Levels of selection, evolution of sex in RNA viruses, and the origin of life. J. theor. Biol. 153, 229–246. C, M. & P, H. H. (1970). Evolution experiments with an artificial ecosystem. J. theor Biol. 28, 393–409. D, R. (1976). The Selfish Gene. Oxford: Oxford University Press.

D, R. (1988) The Blind Watchmaker. Oxford: Oxford University Press. D, D., W, G. E. & F, R. L. (1994). Molecular characterization of catalase from Bordetella pertussis: identification of the katA promoter in an upstream insertion sequence. Mol. Microbiol. 14, 123–130. D, W. F. & S, C. (1980). Selfish genes, the phenotype paradigm and genome evolution. Nature 284, 601–603. E, M. (1985). Macromolecular evolution: dynamical ordering in sequence space. In: Emerging Synthesis in Science: Proceedings of the Founding Workshops of the Sante Fe Institute. (Pines, D. ed.) Sante Fe: Sante Fe Institute. G, R. (1940). The Material Basis of Evolution. New Haven: Yale University Press. G, S. J. & E, N. (1977). Punctuated equilibria: the tempo and mode of evolution reconsidered. Paleobioloy 3, 115–151. G, S. J. & E, N. (1993). Punctuated equilibrium comes of age. Nature 366, 223–227. G, E. P., T, M. L. V., M, M. T., T, J., W, D. J., MF, J. J. et al. (1989). Sequence and characteristics of IS900 , and insertion element identified in a human Crohn’s disease isolate of Mycobacterium paratuberculosis. Nucl. Acid Res. 17, 9063–9073. K, S. (1993). The Origins of Order: Self-Organization and Selection in Evolution. New York: Oxford University Press. K, S. A. & J, S. (1991). Coevolution to the edge of chaos: coupled fitness landscapes, poised states, and coevolutionary avalanches. J. theor. Biol. 149, 467–505. K, N. (1990). Regulation of transposition in bacteria. Annu. Rev. Cell Biol. 6, 297–327. L, S. & R, J. R. (1983). IS200 : a Salmonella-specific insertion sequence. Cell 34, 951–960. N, K., T, H., I, K. & S, T. (1995). Molecular characterization of a tandem repeat, Afa family, and its distribution among Triticeae. Genome 38, 479–486. O, L. E. & C, F. H. C. (1980). Selfish DNA: the ultimate parasite. Nature 284, 604–607. R, M. R. & D, W. F. (1983). Molecular biological mechanisms of speciation. Science 220, 157–162. S, G. G. (1944). Tempo and Mode in Evolution. New York: Columbia University Press. S, A. F., T, G., R, A. D. & J, J. (1995). Ancestral, mammalian-wide subfamilies of LINE-1 repetitive sequences. J. Mol. Biol. 246, 401–417. S, W., C, L. & S, J. G. (1993). The advance of Muller’s ratchet in a haploid asexual population: approximate solutions based on diffusion theory. Genet. Res. 61, 225–231. S, M. (1984). The evolutionary implications of mobile genetic elements. Ann. Rev. Genet. 18, 271–293. T, H. M. (1982). Viruses, protoviruses, development and evolution. Journal of Cellular Biochemistry 19, 105–118. W, C. R., S, E. & L, W. (1985). What are mycoplasmas: the relationship between tempo and mode in bacterial evolution. Journal of Molecular Evolution 21, 305–316. W, C. R. (1987). Bacterial evolution. Microbiol. Rev. 51, 221–271. W, S. (1932). The role of mutation, inbreeding, crossbreeding and selection in evolution. Proceedings of the Sixth International Congress on Genetics 1, 356–366.

APPENDIX The programme CREATURES was written in FORTRAN 77 for the DEC Alpha and compiled using the DEC FORTRAN v6.2 compiler. Each world is initiated with a new population of creatures and a new set of adaptive peaks. The

    programme fills a world array with a population of digital organisms (creatures). Each creature is assigned: an alphabetical character string of n characters (its genome) and has additional registers to record its parental genome, the parent and daughter of its biggest (distance) mutational jump and the fitness difference of that jump, its genome at the last milestone (usually the last printed generation), its mutation rate and its transposon mutation rate as mutations per gene per generation (these are initally set to a global value). Up to nine adaptive peaks, also of n characters are set. Each peak is assigned a quantity of food available on that peak, a regeneration rate that determines the rate of regeneration of food after creature feeding and a pressure for that peak. The programme calculates the distance of each creatures from each peak in turn. However, to model epistatic interactions between genes, the maximum distance between any gene in an epistatic window of K characters and the mapped peak character is used to calculate the grid genetic distance for that window. i.e.: the distance between genome [L-N-T] and peak [F-M-T] is calculated as 18 (3 × 6,—6 being the maximum character distance between L&F, N&M, T&T). To obtain bell-shaped peaks rather than cones, the cosine of the normalised grid distance is used in subsequent calculations. The epistatic window moves along the genome, character by character, and across the ends of the creatures (both creature genomes and peaks modelled as circular strings) and sums the window grid distances for each epistatic window to calculate the creature-peak distance. This is then used to calculate the proximity of the creature to the peak by raising the distance to the power of the pressure for that peak. The creatures are allocated a random place in a queue for obtaining food from that peak. The available food is essentially laid out on a grid and

447

assigned a grid coordinate. Each creature is allocated a random number that maps to the coordinates of the food grid. If a unit of food is located at that coordinate then the creature is allocated that unit. The next cretaure in the queue then takes a turn at obtaining food from the grid. This random element to food allocation is important to ensure that all creatures have a finite chance of obtaining food. Each creature has a number of chances of obtaining food from the grid, allocated according to that creature’s proximity. Creatures that have obtained at least one unit of food may then replicate, producing a daughter for each unit of food they obtained. During replication, the genome of the creatures are allowed to evolve by both micro-mutations that cause a single gene character to shift by a single increment (+ or −) and transposon mutations in which single genes shift by a random increment, at rates determined by each creature’s mutation rates. The mutation rates may be fixed for each creature or may be allowed to vary by fixed increments per generation. After a replication round the parental creatues are lost and a cycle is repeated with the daughter creatures. Parameter’s describing the behaviour of the population may be printed at set generation intervals and at the end of the run. These include global parameters: total population size, mean fitness (average of proximity values for each creature to each peak), average mutation rates, rate of evolution (calculated as the average grid distance moved by each creature since the last milestone); and peak parameters: available food on each peak, closest creature to each peak and its fitness and number of accumulated mutational events, number of creatures (+/− transposon carriers) closest to that peak, average mutation rates of closest creatures.