Modelling Host-Pathogen Co-evolution

Question 1

Modelling host-pathogen co-evolution

Answer 1

1) Matching allele
2) Gene for gene

Question 2

Matching allele framework

Answer 2

• Red Queen (cyclical)
• if you’re a host; the rarer you are the better: negative frequency-dependent selection
• benefit results in an increase in infection; fluctuating selection
• favours polymorphism: everything wants to be rare

Question 3

Gene for gene framework

Answer 3

• arms race (escalatory)
• pathogens of varying levels of generality
• hosts of varying susceptibility
• best to be the host that’s as resistant as possible
• best to be the pathogen that infects the most hosts
• growth cost: a pathogen will need to express different proteins
• can also result in polymorphism as long as there is a cost to improving host range/R

Question 4

Conceptualising 3 loci

Answer 4

• levels of parasite fitness on hosts
• matching allele-type system as a subset of a wider gene-for-gene relationship
• greater gene input, greater complexity output

Question 5

Matching alleles in the real world?

Answer 5

• Daphnia magna versus Pasteuria ramosa
• invertebrate water flea
• stratigraphic analysis: eggs and dormant pathogens from different layers of the pond sediment
• timeshift assay
• “fitness swapping”
• parasites on average seem to be better able to infect contemporary water fleas
• coevolution: matching allele

Question 6

Are noroviruses co-evolving with human blood group antigens - artificial expt?

Answer 6

• Virus Like Particles based on
  the GII.4 norovirus Hunter
  strain
• binding preferences to different
  glycans from different blood
  groups
• Wt: good at binding to secretor glycans
• introduce artificial aa mutations into ORF2 viral capsid protein: becomes able to infect non-secretors
• changing host range by changing protein
• recapped by both models and experimental data

Question 7

Are noroviruses co-evolving with human blood group antigens - genetic expt?

Answer 7

• changes over time: lots of evolution is happening
• different strains succeed at different points in time; dynamic system
  H: precursor
• H gets modified to become A or B

Question 8

Are P. falciparum proteins co-evolving with sickle cell?

Answer 8

• amino acid changes in P. falciparum
• small fraction of parasite strains
• functional, gene-for-gene-like changes seen at a higher level with higher frequencies of sickle cell

Question 9

Are MHC molecules co-evolving with pathogens?

Answer 9

• we have more than one gene
• we need to consider multiple genes coevolving

Question 10

Immune selection can shape the population structure of pathogens

Answer 10

• things pathogen is displaying that adjust can become immune targets
• epitopes
• exhibit strain structure
• cross-immunity between strains is determined by pathogens which do not share antigenic determinants
• these can co-circulate

Question 11

A model of immune selection shaping the population structure of pathogens - the basics

Answer 11

• host infected by different strains
• depending on host genotype, two outcomes

Question 12

A model of immune selection shaping the population structure of pathogens - the specifics

Answer 12

1) pathogen adopts a permanent non-overlapping strain structure
2) pathogen strain structure cycles from one state to another

Question 13

What if this structure creates differential selection pressures on the host?

Answer 13

• hosts who can recognise most epitopes
• hosts who can only recognise a
  few epitopes

Question 14

hosts who can recognise most epitopes

Answer 14

force pathogens to exist as
immunologically discrete strains

Question 15

hosts who can only recognise a
few epitopes

Answer 15

• determine which strain structure
  the pathogen adopts
• differentially killed by particular pathogen strains

Question 16

permanent, non-overlapping strain structure

Answer 16

• some strains come to dominate
• hosts that can’t deal just get driven out
• MHC types are lost
• drives the host recognition alleles into a state of complete linkage disequilibrium
• creates a a fixed state
• pathogen imposes some structure

Question 17

Non-overlapping associations between MHC loci

Answer 17

• could be the result of pathogen selection

Question 18

pathogen strain structure cycles from one state to another

Answer 18

• cycle between combinations of non-overlapping strains dominate in different points in time
• structuring amoung the host recognition alleles cycle too
• drives MHC types to be more or less advantageous at different points in time

Question 19

Example of pathogen selection resulting in non-overlapping associations between MHC loci: Pakistan

Answer 19

• HLA-A/B alleles emerge in population genomes
• signature of pathogen selection, explained by interactions

Question 20

Example of pathogen selection resulting in non-overlapping associations between MHC loci: Mayans

Answer 20

• Ancient Mayan population in Mexico: pre-massive pathogen (S. typhimurium) selection event
• before, it would have co-evolved
• Columbian era of Europeans coming to the Americas
• population bottlenecks
• contrast to a modern population MHCs
• striking: F measure
• modern pop exhibits much more extreme non-overlap between HLAs than ancient pops

Question 21

ancient DNA studies

Answer 21

• technology
• circumstantial; hard to prove

Question 22

F-measure

Answer 22

• measure of non-overlap
• is non-overlap extreme relative to a random case?

Question 23

How do we identify infection associated host loci?

Answer 23

1) case control studies, examining candidate loci
2) GWAS

Question 24

HLA case control study

Answer 24

• odds ratio
• 0.42 protection

Question 25

odds ratio

Answer 25

which is the ratio of the odds of having the disease outcome in the exposed group, to the odds of having the disease outcome in the non exposed group.

Question 26

GWAS

Answer 26

- sequence cases and controls - looking at a whole genome - genomic SNP library: scattered; not candidates but within candidate regions - calculate stats: are any of the SNPs associated with having the disease? - no candidate locus - one association is ideal - stats problem: multiple comparisons; threshold of the p value to be deemed significant is very stringent

Question 27

SNP library

Answer 27

we cant look at every nt

Question 28

How does GWAS work?

Answer 28

- the association will appear due to linkage disequilibrium between the SNPs that have been genotyped and the mutation that has the effect of interest - choice of SNP array really matters: a GWAS needs to include appropriate SNPs for the population being studied (they are pop-specific)

Question 29

GWAS: severe malaria

Answer 29

- glycophorin

Question 30

Complexity of dealing with potentially co-evolving systems

Answer 30

- population frequency of an MHC variant which recognises one of the pathogen strains - odds ratio for the odds of being infected if you have the given MHC variant - case control studies in silico see associations all the time, because of the many strains, and many MHC variants, in the system