Molecular Epidemiology

1
Molecular Epidemiology
2

• This is the principle technique of scientific
  inquiry by changing the scale of description, we
  move from unpredictable, unrepeatable individual
  cases to collections of cases whose behavior is
  regular enough to allow generalizations to be
  made. (S. Levin, 1947)

3
Epidemiology

• Originally Study(ology) upon(Epi) populations of
  people(demes)
• Now much broader.
• inquiry into events that take place over very
  different temporal scales From identification
  of organisms that have diverged millions of years
  ago, to the tracing of contacts.

4
On a Large Scale
Identity of an infectious agent in an outbreak
5
Ribosomal RNA
Coding Regions Highly conserved across widely
divergent species. Transcribed Spacer Regions
Less conserved. Different between closely
related Species. Non-transcribed Spacer
Regions Vary between and among species.
6
(No Transcript)
7
Where is it?
8
What is it?

• Microscopy
• Molecular Methods

Figure 1.-Oocysts of a Cyclospora Species (Panel
A), Cryptosporidium muris (Panel B), and C.
parvum (Panel C) (Modified Acid-Fast Stain)
DNA sequencing. Use Moderately variable
regions. Such as the transcribed spacer.
9
Cyclospora ITS-1

• Use conserved primers from the flanking (coding
  regions),
• amplify and sequence ITS.
• 2. Design primers common to all Cyclospora
  isolates.
• 3. Test sensitivity and specificity of primers.

• 1. Use conserved primers from the flanking
  (coding regions), amplify and sequence ITS.
• 2. Design primers common to all Cyclospora
  isolates.
• 3. Test sensitivity and specificity of primers.

Amplified 36 C. cayetanensis from around the
world Did NOT amplify 20 species with similar
pathology Among them Cryptosporidia. Faint band
from Babesia gibsoni.
10
Assumptions

• False Positives Less stringent PCR conditions
• False negatives Overly stringent conditions,
• combined with unforeseen
• mutation in primer regions

11
Zooming in

• The way to study events on a large scale may not
  be the way to study events on a small scale
  (think physics)
• What is TRUE on one scale may not be true on
  another scale.

12
On a Smaller Scale
Strains Transmission cycles
13
GIARDIA
Giardia has Two Heads!
14
Mycobacterium tuberculosis

• According to the WHO

15
Mycobacterium tuberculosis

• According to the WHO
• 2 Billion infected

16
Mycobacterium tuberculosis

• According to the WHO
• 2 Billion infected
• 1/10 will become sick

17
Mycobacterium tuberculosis

• According to the WHO
• 2 Billion infected
• 1/10 will become sick
• 2.7 million die each year

18
Mycobacterium tuberculosis

• According to the WHO
• 2 Billion infected
• 1/10 will become sick
• 2.7 million die each year
• TB is the largest single agent killer of

19
Mycobacterium tuberculosis

• According to the WHO
• 2 Billion infected
• 1/10 will become sick
• 2.7 million die each year
• TB is the largest single agent killer of Women.

20
Mycobacterium tuberculosis

• According to the WHO
• 2 Billion infected
• 1/10 will become sick
• 2.7 million die each year
• TB is the largest single agent killer of Women.
  Young.

21
Mycobacterium tuberculosis

• What is the frequency of exogenous re-infection?
  With
• MDR-TB?
• What are the transmission dynamics in endemic
• countries?

22
Methods to differentiate strains

• Isoenzymes/allozymes older methods.
• RFLP
• RAPD/ AP-PCR
• AFLPs
• Sequence surrogates report nucleotic changes
  indirectly

23
Isoenzymes

• Isoenzymes/allozymes electrophoresis to
  determine differences in enzymes. Allozymes
  detect differences between alleles of a given
  enzyme. Very weak.
• Detect 60 of change, only at enzyme loci.
• Giardia divided into 2 clades evidence for
  zoonosis

24
RFLP

• Restriction fragment length polymorphism
• Usually a true sequence surrogatea difference in
  RFLP pattern is ideally due to a change in the
  nucleotide sequence at one or many restriction
  sites.
• RFLPs are highly dependent on experimental
  conditions.

25
GIARDIA RFLP of Intergenic rRNA Spacer (IGS)
RFLP of the IGS locus differentiates Four strains
compared to 2 identified By isoenzyme analysis.
26
TB-RFLP with Insertion Sequences

• IS6110- Fingerprinting use alu to digest
  genome. Little variation in RFLP. Question is,
  in which fragments is the insertion element
  present?
• IS6110 is a transposon that jumps around the
  genome.
• IS6110 is not purely a sequence surrogate, it
  is also a transposon surrogate

27
IS6110

• The ruler is ALIVE
• It is dynamic, and reaches equilibrium slower
  than TB in an outbreak.

28
IS6110

• of IS6110 copies in TB genomes varies from 0 to
  25. When copy number is low, klt5, there is less
  change in fingerprints
• -contact investigation is very hard.

29
RAPD or AP-PRC

• RAPD/AP-PCR- Amplify with random primers.
• Sequence surrogateTests whether there is a
  change in the template regions only. Analysis is
  the same as that for RFLP.
• Cycles of low-stringency leads to amplification
  of contaminants.
• Highly dependent on reaction conditions.
• Groupings correspond to Isozymes.

30
AFLPs

• AFLPs digest DNA, ligate to adaptors, PCR
• Dont need low-stringency steps, less
  non-specific amplification.
• Same analysis as RFLPs, need .2 to 1mg of DNA.
• No good for Giardia and other parasitesneed too
  much DNA.

31
Smaller Still
Identifying Clonal Lineages Tracking transmission
32
Methods

• Minisatellites
• Microsatellites
• IGS rDNA intergenomic spacer

33
Microsatellites

• Simple Sequence Repeats
• Repeating motifs for 2-5bp
• Scattered throughout the genome
• Amenable to PCR and cloning due to small allele
  size.

34
Minisatellites

• Repeating motifs 10-100 bp
• Analysed with DNA
• probes specific for a single locus.

35
TB Spoligotyping

• Spacer Oligotyping
• Direct repeat (DR) locus 36bp, freq. varies
• Use primers somewhere in the DR, amplify
  non-repetitive spacer sequences 34-41bp
• Identify the spacers by hybridization to know
  sequence oligonucleotides
• Need sequence to generate the oligos

36
Depends on

• Dynamics of DR regions.
• Change in sequence in non-repetitive regions.
• DR regions-are they at equilibrium?
• How often do they repeat?
• -Not yet known

37
Spoligotypes vs. IS6110

• IS6110 IS6110 types Spoligotypes
• 1 1 10
• 2-5 7 8
• gt5 80 52
• Spoligotyping can identify M. bovis (BCG vaccine)
• Detection and strain differentiation can be done
• Simultaneously without culture.

38
Crossing scales
39
Crossing scales

• DNA sequence of small subunit (SSU) ribosomal RNA
  (highly conserved) suggests four groups of
  Giardia. Groups 3 and 4 are only in Dogs. 293bp
• 1-------GCG------_G---------T-------C-------------
  ------
• 2-------ATC-------AC---------G------G-------------
  ------
• 3-------ATC-------AC---------A------G---------T---
  -----
• 4-------ATC-------AC---------A------A----------T--
  --A-
• 1 and 2 are mainly in humans, though some dogs
  have 3. 2,3,4 and four are nearly identical
• Is this good evidence against zoonosis?

40
Models of Nucleotide Substitution

• On a large scale, we can calculate the rate of
  substitution, then estimate the likelihood of any
  given substitution and control for confounders
  (transition-transversion, codon bias etc).
• On a small scale we do not know rate, the process
  is nearly random, and confounders may be
  irrelevant

41
Distributions
BINOMIAL Pr(Yy)n!/(y!(n-y) Py(1-P)n-y
Mean nP Variance nP(1-P) POISSON Pr(Yy)
uye-u / y! Mean and Variance u Central Limit
Theorem Large number of events? normal
distribution Binomial- coin toss. Poisson- rare
events. Tossing a 100,000 sided die.
42
Kimuras 2 parameter

• For instance, as the rate of transition and
  transversion become small Kimuras 2 parameter
  model reduces to a one parameter model
• K -(1/2) ln1-2P-Qv(1-2Q) ?
• KP Q
• where K is the distance per site
• and P and Q are the fractions of sites with
  transition vs/ transversion changes.

43
How to Analyze RFLP and other sequence surrogates

• Two sources of information number of bands, and
  size of each fragment.
• -In practice, it can be difficult to score
  changes in fragment size. Most studies
• look only at the presence or absence of
• a certain pattern.

44
Nei and Lis model for RFLP

• The expected frequency of restriction sites with
  r nucleotide pairs depends on GC content and GC
  content of restriction site sequence
• A (g/2)r1(1-g)/2r2
• G GC of genome
• r1, r2 are GC, and AT frequencies in
  Restriction site. r1r2r

45

• mtnumber of nucleotide pairs in genome
• mta n, the expected of restriction sites
• What is the probability that the n changes over
  time t?

46

• Mutations are a Poisson process.
• P e-rlt
• lMutation rate/nucleotide
• r Length of restriction sequence
• t Time

47
Nei and Li continued

• n(t) number of bands at time t n1(t) n2(t)
• n1(t) of sites that do not change
• n2(t) number of new sites.
• E(n)n0P mta(1-P) or E(n2) E(n1)
• Variance n1(t) and n2(t) are independent
• Var n(t) Varn1(t)Varn2(t)
• n1(t) is binomial, n2(t) is poisson
• Var n(t)n0P(1-P) mta(1-P)

48
IS6110 is modelled similarily

• Transposition is raremodeled as a Poisson
  process
• Prob of at least 1 change 1-ekqt
• Where k of copies of transposon in genome
• And q is the rate of transposition when k1

49
Really Small-New Technology

• Genetic marking of drug resistance, or virulence
• -Represenational Difference Analysis (RDA)
• -High-throughput genotyping
• -Microarrays

50
Representational Difference Analysis

• Cloning the Differences Between Two Complex
  Genomes Lisitsyn Science, feb 1993
• Uses Subtractive and Kinetic enrichment to purify
  fragments present in one population, but absent
  in another.
• Basically differential amplification of
  polymorphic fragments

51
High-Throughput Genotyping

• Flourescent labels incorporated into RAPDs,
  microsatellites and AFLP
• Can run in ONE electrophoresis lane.
• Result complicated fingerprints that take into
  account variation at different levels.

52
Conclusions

• 1 The strongest analyses will be those that
  consider variation on multiple temporal levels.
• 2. Everyone says their technique is economically
  feasible for use in endemic countries no one
  says how much their technique costs.
• 3. Stay away from Guatemalan raspberries.