This repository houses work done as part of the project and publication:
"Evolutionary genetics of immunological supertypes reveals two faces of the Red Queen"
by
- Jackie Lighten
- Alexander S.T. Papadopulos
- Ben J. Ward
- Ian Paterson
- Lyndsey Baillie
- Ian R. Bradbury
- Andrew P. Hendry
- Paul Bentzen
- Cock van Oosterhout
DOI: doi:10.1038/s41467-017-01183-2
Two parts of the project are included in this repository:
1). Computation of Jost D on Super-Type data, and bootstrapping simulations.
This repository houses scripts used for the computation of a measure of differentiation called JostD, on the sample data. In addition, random bootstrap simulations were performed, with the aim of randomizing supertype category allocations to sequenced alleles. To see if JostD estimated from the actual data differed significantly from the distribution of estimates of JostD computed for randomized simulations.
Much of the computation of JostD and the randomized simulations is handled by a package used in these scripts called SuperTypeR, so as other users wanting to use similar calculations in the future can include SuperTypeR in their own scripts.
SuperTypeR, and the scripts in this Repo were written by co-author Ben J. Ward (Github: Ward9250).
2). Simulations of immune gene evolution in a host-parasite system.
The code provided as .mac files in the src folder is for a Minitab 12.1 macro to study the evolution of immune genes in a host-parasite system, examining whether trans-species polymorphisms of STs can evolve in a Red Queen's arms race.
In particular, it analyses the adaptive evolutionary change in epitope recognition of immune alleles and STs (i.e. their paratope) during host parasite co-evolution using an agent-based model.
The Minitab 12.1 macro is for the most basic model that (1) would result in antagonistic host-parasite co-evolution, and (2) in which we could quantify the resulting adaptive evolutionary change in phenotype over time.
Hence, rather than using a strict population genetic model, it models the paratope of immune alleles and the epitope of parasites in a 2-D grid with size 1000 x 1000.
The fitness effect of immune alleles was determined by the position of immune alleles and parasites in the epitope/paratope space. The adaptive evolutionary change in phenotype of alleles and STs was quantified by tracking changes in their position within this space over time. Analysing the phenotypic change enabled us to study trans species evolution. Furthermore, by analysing fluctuations in immune allele frequencies, we were able to study the population genetic characteristics of the model.
Hosts were diploid with one immune locus. Parasites were haploid. Each host was infected by one parasite every generation. The minimum Euclidean distance was calculated between an individual’s immune alleles and one randomly drawn parasite representing the infection. Depending on this distance, the parasite was either recognized (in a resistant host) or not (in a susceptible host). Fitness was relative so that 50% of all parasites died (on resistant hosts). The other 50% of parasites (on susceptible hosts) reproduced clonally one individual offspring. Parasite offspring mutated, causing them to change their X or Y coordinates by one unit within the grid. Parasite infection on the susceptible hosts reduced host fitness by 0.25, and host with zero fitness died. Resistant host gained 0.25 fitness units, and individuals with one fitness units reproduced offspring that all started with 0.25 fitness units.
Reproduction of hosts was sexual, and hosts produced gametes, each containing one parental immune allele. This immune allele could mutate with probability µ, which caused it to change its X or Y coordinates by one unit within the grid. Note that by simulating a high mutation rate, we effectively accelerated evolutionary time in the model.
For example, simulations with µ= 0.1 equate to 3.1 x 106 generations in real time, assuming a base mutation rate of 10-9 per base per generation, and 16 PBR codons with a total of 32 replacement sites (i.e. the 1st and 2nd codon positions of the PBR). Gametes of reproducing hosts united randomly with one another to produce the next generation of diploid offspring. This resulted in a Poisson distribution of offspring per parent (mean=variance=unity). Output is displayed in the Minitab 12.1 Data Window.
3). Plotting scripts
We also tried plotting Supertype evolution using R with different colour schemes for publication. Some of these plots, data used to make them, and the scripts are included in this repository.
- data - Raw data given to me at start of project or computational experiments.
- results - The result files of computational experiments.
- src - Source code of binary executable files used in this project and / or scripts.
- R with the SuperTypeR package installed.
- analysis: Calculate JostD of the data, and do bootstrap simulations of JostD.
File: SuperType_Designations.csv
This contains the original supertype designation for each allele. The first column is the supertype, the second is genotype, and the third is the sequence of the genotype.
File: Samples.csv
This contains the genotypes per individual per population and the corresponding STs. Each row is one sample.
Columns are (left to right):
- Sample ID
- Population
- Drainage
4-12. Alleles in sample
14-22. Supertypes corresponding to alleles in columns 4-12. - Number of alleles in sample.
- Number of supertypes in sample.
Files: The .xls files, and ST_RQ_Dynamics.csv
These contain the data used for plotting the plots in results/plotting/
The script in src/DataAnalysis.R scripts the the analysis for this project.
It pre-processes the input data file data/Samples.csv, before using the
SuperTypeR package to compute JostD
on the data in Samples.csv, before doing bootstrap simulations to estimate the
mean and standard error of the JostD estimates.
Supertype counting, computation of of JostD and Pairwise JostD is done by the SuperTypeR package.
The aim is to randomly assign the alleles to the supertypes in file SuperType_Designations.csv, generating a set of artificial supertypes composed of a random set of alleles. The analysis can then be re-run on these null datasets to assess whether the results from the actual data are significantly different.
FOR EACH IN N ITERATIONS:
-
Wipe ST designations across all samples in data from Samples.csv.
-
For the data in SuperType_Designations.csv, randomly reallocate the alleles in the total gene pool (~539) to the supertypes (the supertypes still get the same number of alleles allocated to them). Allocation is done without replacement.
-
For each population in Samples.csv, count the number of each supertype present in the population. This results in a matrix of supertype counts per population, where each row is a supertype, and each column is a population.
-
Calculate the pairwise Jost D for the matrix output from step 3. Also calculate the average of each column i.e. the mean Jost D for each population relative to all others.
The julia and R scripts called RQ_Evo_Plotting.R, and RE_Evo_Plotting.jl in the src/ directory are were used to process the data in ST_RQ_Dynamics.csv (which was saved from the .xls files), to make the plots in results/plotting.