HIV epidemiology Routes of infection MSM: Men who have sex with men - - PowerPoint PPT Presentation

▶
hiv epidemiology
SMART_READER_LITE
LIVE PREVIEW

HIV epidemiology Routes of infection MSM: Men who have sex with men - - PowerPoint PPT Presentation

Mar 06, 2023 564 likes •1.2k views

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi Max Planck Institute for Informatics AREVIR, May 8 2015 eu resist HIV epidemiology Routes of infection MSM: Men who have sex with men Phylogeny-based HIV transmission networks

slide-1
SLIDE 1

Phylogeny-based HIV transmission networks

Prabhav Kalaghatgi

Max Planck Institute for Informatics

AREVIR, May 8 2015

euresist

slide-2
SLIDE 2

HIV epidemiology

Routes of infection MSM: Men who have sex with men

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 2/18

slide-3
SLIDE 3

HIV epidemiology

Routes of infection MSM: Men who have sex with men HET: Heterosexual partners

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 2/18

slide-4
SLIDE 4

HIV epidemiology

Routes of infection MSM: Men who have sex with men HET: Heterosexual partners IDU: Injection drug users

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 2/18

slide-5
SLIDE 5

HIV epidemiology

Routes of infection MSM: Men who have sex with men HET: Heterosexual partners IDU: Injection drug users

Motivation

A better understanding of the HIV epidemic through the analysis

  • f person-to-person transmission network.

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 2/18

slide-6
SLIDE 6

Rapidly evolving pathogens

Accumulation of genomic mutations via a series of transmissions

1 2 3 4 5 6

1 ATAGGTCCATAGCCAGATTGGCCAAATAGATCCACCAGATTGGCCACCATAC

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 3/18

slide-7
SLIDE 7

Rapidly evolving pathogens

Accumulation of genomic mutations via a series of transmissions

1 2 3 4 5 6

1 ATAGGTCCATAGCCAGATTGGCCAAATAGATCCACCAGATTGGCCACCATAC 2 ATAGGTCCATAGCCAGATTGCCCAAATAGATCCACCAGATTGGCCACCATAC

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 3/18

slide-8
SLIDE 8

Rapidly evolving pathogens

Accumulation of genomic mutations via a series of transmissions

1 2 3 4 6

1 ATAGGTCCATAGCCAGATTGGCCAAATAGATCCACCAGATTGGCCACCATAC 2 ATAGGTCCATAGCCAGATTGCCCAAATAGATCCACCAGATTGGCCACCATAC 3 ATAGGTCCATAGCCAGATTGCCCAAATAGATCCACCAGACTGGCCACCATAC

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 3/18

slide-9
SLIDE 9

Rapidly evolving pathogens

Accumulation of genomic mutations via a series of transmissions

1 2 3 4

1 ATAGGTCCATAGCCAGATTGGCCAAATAGATCCACCAGATTGGCCACCATAC 2 ATAGGTCCATAGCCAGATTGCCCAAATAGATCCACCAGATTGGCCACCATAC 3 ATAGGTCCATAGCCAGATTGCCCAAATAGATCCACCAGACTGGCCACCATAC 4 ATAGGTCCATAGCCAGATTGCCCAAATAGAACCGCCAGATTGGCCACCATAC

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 3/18

slide-10
SLIDE 10

Rapidly evolving pathogens

Accumulation of genomic mutations via a series of transmissions

1 2 3 4 5

1 ATAGGTCCATAGCCAGATTGGCCAAATAGATCCACCAGATTGGCCACCATAC 2 ATAGGTCCATAGCCAGATTGCCCAAATAGATCCACCAGATTGGCCACCATAC 3 ATAGGTCCATAGCCAGATTGCCCAAATAGATCCACCAGACTGGCCACCATAC 4 ATAGGTCCATAGCCAGATTGCCCAAATAGAACCGCCAGATTGGCCACCATAC 5 ATAGGTCCATAGCCAGATTGCCCAAATGGATCCACCAGACTGGCCACCATAC

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 3/18

slide-11
SLIDE 11

Rapidly evolving pathogens

Accumulation of genomic mutations via a series of transmissions

1 2 3 4 5 6

1 ATAGGTCCATAGCCAGATTGGCCAAATAGATCCACCAGATTGGCCACCATAC 2 ATAGGTCCATAGCCAGATTGCCCAAATAGATCCACCAGATTGGCCACCATAC 3 ATAGGTCCATAGCCAGATTGCCCAAATAGATCCACCAGACTGGCCACCATAC 4 ATAGGTCCATAGCCAGATTGCCCAAATAGAACCGCCAGATTGGCCACCATAC 5 ATAGGTCCATAGCCAGATTGCCCAAATGGATCCACCAGACTGGCCACCATAC 6 ATAGATCCATAGCCAGATTGCCCAAATAGAACCGCCAGATTGCCCACCATAC

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 3/18

slide-12
SLIDE 12

Objective

Recover transmission network from sequences sampled from infected individuals

1 ATAGGTCCATAGCCAGATTGGCCAAATAGATCCACCAGATTGGCCACCATAC 2 ATAGGTCCATAGCCAGATTGCCCAAATAGATCCACCAGATTGGCCACCATAC 3 ATAGGTCCATAGCCAGATTGCCCAAATAGATCCACCAGACTGGCCACCATAC 4 ATAGGTCCATAGCCAGATTGCCCAAATAGAACCGCCAGATTGGCCACCATAC 5 ATAGGTCCATAGCCAGATTGCCCAAATGGATCCACCAGACTGGCCACCATAC 6 ATAGATCCATAGCCAGATTGCCCAAATAGAACCGCCAGATTGCCCACCATAC

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 4/18

slide-13
SLIDE 13

Objective

Recover transmission network from sequences sampled from infected individuals

1 ATAGGTCCATAGCCAGATTGGCCAAATAGATCCACCAGATTGGCCACCATAC 2 ATAGGTCCATAGCCAGATTGCCCAAATAGATCCACCAGATTGGCCACCATAC 3 ATAGGTCCATAGCCAGATTGCCCAAATAGATCCACCAGACTGGCCACCATAC 4 ATAGGTCCATAGCCAGATTGCCCAAATAGAACCGCCAGATTGGCCACCATAC 5 ATAGGTCCATAGCCAGATTGCCCAAATGGATCCACCAGACTGGCCACCATAC 6 ATAGATCCATAGCCAGATTGCCCAAATAGAACCGCCAGATTGCCCACCATAC

1 2 3 4 5 6

Fully resolved network

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 4/18

slide-14
SLIDE 14

Objective

Recover transmission network from sequences sampled from infected individuals

1 ATAGGTCCATAGCCAGATTGGCCAAATAGATCCACCAGATTGGCCACCATAC 2 ATAGGTCCATAGCCAGATTGCCCAAATAGATCCACCAGATTGGCCACCATAC 3 ATAGGTCCATAGCCAGATTGCCCAAATAGATCCACCAGACTGGCCACCATAC 4 ATAGGTCCATAGCCAGATTGCCCAAATAGAACCGCCAGATTGGCCACCATAC 5 ATAGGTCCATAGCCAGATTGCCCAAATGGATCCACCAGACTGGCCACCATAC 6 ATAGATCCATAGCCAGATTGCCCAAATAGAACCGCCAGATTGCCCACCATAC

1 2 3 4 5 6

Fully resolved network infeasible

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 4/18

slide-15
SLIDE 15

Objective

Recover transmission network from sequences sampled from infected individuals

1 ATAGGTCCATAGCCAGATTGGCCAAATAGATCCACCAGATTGGCCACCATAC 2 ATAGGTCCATAGCCAGATTGCCCAAATAGATCCACCAGATTGGCCACCATAC 3 ATAGGTCCATAGCCAGATTGCCCAAATAGATCCACCAGACTGGCCACCATAC 4 ATAGGTCCATAGCCAGATTGCCCAAATAGAACCGCCAGATTGGCCACCATAC 5 ATAGGTCCATAGCCAGATTGCCCAAATGGATCCACCAGACTGGCCACCATAC 6 ATAGATCCATAGCCAGATTGCCCAAATAGAACCGCCAGATTGCCCACCATAC

1 2 3 4 5 6

Fully resolved network

1 2 3 4 5 6

Partially resolved network infeasible

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 4/18

slide-16
SLIDE 16

HIV sequence data: EuResist

Selection criteria

Subtype B Country of origin in Europe Oldest sequence per individual Sampling times from 1999 to 2012

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 5/18

slide-17
SLIDE 17

HIV sequence data: EuResist

Selection criteria

Subtype B Country of origin in Europe Oldest sequence per individual Sampling times from 1999 to 2012

Data used

15,000 sequences Transmission mode: MSM (25%), HET (22%), IDU (19%), Unknown (34%)

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 5/18

slide-18
SLIDE 18

Outline

Threshold-based networks

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 6/18

slide-19
SLIDE 19

Outline

Threshold-based networks Threshold-free networks using phylogenetic distances

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 6/18

slide-20
SLIDE 20

Outline

Threshold-based networks Threshold-free networks using phylogenetic distances Timed networks using molecular clock

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 6/18

slide-21
SLIDE 21

Europe-wide transmission networks

Network constructed by thresholding distance between sequences (LogDet)

Kalaghatgi et al. European Workshop on HIV & HCV 2013 Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 7/18

slide-22
SLIDE 22

Europe-wide transmission networks

Network constructed by thresholding distance between sequences (LogDet) Low cross-country transmission

Kalaghatgi et al. European Workshop on HIV & HCV 2013 Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 7/18

slide-23
SLIDE 23

Europe-wide transmission networks

Network constructed by thresholding distance between sequences (LogDet) Low cross-country transmission Small cluster sizes

Kalaghatgi et al. European Workshop on HIV & HCV 2013 Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 7/18

slide-24
SLIDE 24

Europe-wide transmission networks

Network constructed by thresholding distance between sequences (LogDet) Low cross-country transmission Small cluster sizes 25% of sequences are linked

Kalaghatgi et al. European Workshop on HIV & HCV 2013 Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 7/18

slide-25
SLIDE 25

Accurate (tree-based) distances

LogDet, TN93, Hamming distance yield inaccurate estimates

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 8/18

slide-26
SLIDE 26

Accurate (tree-based) distances

LogDet, TN93, Hamming distance yield inaccurate estimates Model-based estimates are more accurate

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 8/18

slide-27
SLIDE 27

Accurate (tree-based) distances

LogDet, TN93, Hamming distance yield inaccurate estimates Model-based estimates are more accurate – Phylogenetic tree

Phylogenetic tree a b c d e f

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 8/18

slide-28
SLIDE 28

Accurate (tree-based) distances

LogDet, TN93, Hamming distance yield inaccurate estimates Model-based estimates are more accurate – Phylogenetic tree – Substitution models:

GTR, HKY

Phylogenetic tree a b c d e f

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 8/18

slide-29
SLIDE 29

Accurate (tree-based) distances

LogDet, TN93, Hamming distance yield inaccurate estimates Model-based estimates are more accurate – Phylogenetic tree – Substitution models:

GTR, HKY

– Among-site rate

variation model

Phylogenetic tree a b c d e f

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 8/18

slide-30
SLIDE 30

Accurate (tree-based) distances

LogDet, TN93, Hamming distance yield inaccurate estimates Model-based estimates are more accurate – Phylogenetic tree – Substitution models:

GTR, HKY

– Among-site rate

variation model

Optimize models and phylogenetic tree

Phylogenetic tree a b c d e f

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 8/18

slide-31
SLIDE 31

Threshold-free networks

a b c d e f

Phylogenetic tree

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 9/18

slide-32
SLIDE 32

Threshold-free networks

a b c d e f

Phylogenetic tree

d c b a e f a b c d e f

Tree-based distances

3.1 4.1 5.9 2.6 5.3 3.4 9.8 6.4 7.5 8.5 4.2 5.8 8.6 6.5 6.1 _ _ _ _ _ _

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 9/18

slide-33
SLIDE 33

Threshold-free networks

a b c d e f

Phylogenetic tree

d c b a e f a b c d e f

Tree-based distances

3.1 4.1 5.9 2.6 5.3 3.4 9.8 6.4 7.5 8.5 4.2 5.8 8.6 6.5 6.1 _ _ _ _ _ _

Objective: Among all undirected trees select one that minimizes the sum of edge weights (distances)

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 9/18

slide-34
SLIDE 34

Threshold-free networks

a b c d e f f e c b a

Phylogenetic tree

d

Transmission network

d c b a e f a b c d e f

Tree-based distances

3.1 4.1 5.9 2.6 5.3 3.4 9.8 6.4 7.5 8.5 4.2 5.8 8.6 6.5 6.1 _ _ _ _ _ _

Objective: Among all undirected trees select one that minimizes the sum of edge weights (distances) Optimal tree is the minimum spanning tree.

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 9/18

slide-35
SLIDE 35

Heuristics for handling large data

  • 1. Approximate
  • ptimization

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 10/18

slide-36
SLIDE 36

Heuristics for handling large data

  • 1. Approximate
  • ptimization
  • 2. Accurate optimization
  • f constrained trees

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 10/18

slide-37
SLIDE 37

Heuristics for handling large data

  • 1. Approximate
  • ptimization
  • 2. Accurate optimization
  • f constrained trees
  • 3. Network construction

using minimum spanning tree

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 10/18

slide-38
SLIDE 38

Heuristics for handling large data

  • 1. Approximate
  • ptimization
  • 2. Accurate optimization
  • f constrained trees
  • 3. Network construction

using minimum spanning tree

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 10/18

slide-39
SLIDE 39

Consensus transmission networks

  • ●
  • ●
  • ●
  • ●
  • Edge

support 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 11/18

slide-40
SLIDE 40

Network size vs robustness

Edge support: proportion of networks that contain the edge

  • 10

20 30 40 0.0 0.2 0.4 0.6 0.8 1.0

Transmission network size Edge support

1 1 1 2 1 2 2 1 2 2 3 2 3 4 7 4 8 17 23 42 50 87 221

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 12/18

slide-41
SLIDE 41

Topological correspondence

a b c d e f f e c b a

Phylogenetic tree

d

Transmission network

Each subtree in the phylogenetic tree induces a connected subgraph in the transmission network

Kalaghatgi et al. Bioinformatics; under review Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 13/18

slide-42
SLIDE 42

Topological correspondence

a b c d e f f e c b a

Phylogenetic tree

d

Transmission network

Each subtree in the phylogenetic tree induces a connected subgraph in the transmission network

Kalaghatgi et al. Bioinformatics; under review Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 13/18

slide-43
SLIDE 43

Topological correspondence

a b c d e f f e c b a

Phylogenetic tree

d

Transmission network

Each subtree in the phylogenetic tree induces a connected subgraph in the transmission network

Kalaghatgi et al. Bioinformatics; under review Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 13/18

slide-44
SLIDE 44

Topological correspondence

a b c d e f f e c b a

Phylogenetic tree

d

Transmission network

Each subtree in the phylogenetic tree induces a connected subgraph in the transmission network

Kalaghatgi et al. Bioinformatics; under review Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 13/18

slide-45
SLIDE 45

Topological correspondence

a b c d e f f e c b a

Phylogenetic tree

d

Transmission network

Each subtree in the phylogenetic tree induces a connected subgraph in the transmission network Low support in large networks due to difficulty in resolving distant divergence events

Kalaghatgi et al. Bioinformatics; under review Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 13/18

slide-46
SLIDE 46

Constructing timed networks

Kalaghatgi et al. European Workshop on HIV & HCV 2014 Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 14/18

slide-47
SLIDE 47

Constructing timed networks

So far we constructed untimed networks

Kalaghatgi et al. European Workshop on HIV & HCV 2014 Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 14/18

slide-48
SLIDE 48

Constructing timed networks

So far we constructed untimed networks Sampling times allow an estimation

  • f transmission time

Kalaghatgi et al. European Workshop on HIV & HCV 2014 Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 14/18

slide-49
SLIDE 49

Constructing timed networks

So far we constructed untimed networks Sampling times allow an estimation

  • f transmission time

A B transmission time divergence time sampling time A sampling time B time virus lineage in A virus lineage in B

Kalaghatgi et al. European Workshop on HIV & HCV 2014 Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 14/18

slide-50
SLIDE 50

Constructing timed networks

So far we constructed untimed networks Sampling times allow an estimation

  • f transmission time

A B transmission time divergence time sampling time A sampling time B time virus lineage in A virus lineage in B

Convert evolutionary distance to time using molecular clock

Kalaghatgi et al. European Workshop on HIV & HCV 2014 Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 14/18

slide-51
SLIDE 51

Constructing timed networks

So far we constructed untimed networks Sampling times allow an estimation

  • f transmission time

A B transmission time divergence time sampling time A sampling time B time virus lineage in A virus lineage in B

Convert evolutionary distance to time using molecular clock Calibrate molecular clock using sampling times

Kalaghatgi et al. European Workshop on HIV & HCV 2014 Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 14/18

slide-52
SLIDE 52

Constructing timed networks

So far we constructed untimed networks Sampling times allow an estimation

  • f transmission time

A B transmission time divergence time sampling time A sampling time B time virus lineage in A virus lineage in B

Convert evolutionary distance to time using molecular clock Calibrate molecular clock using sampling times Construct transmission networks with edges labeled with transmission time

Kalaghatgi et al. European Workshop on HIV & HCV 2014 Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 14/18

slide-53
SLIDE 53

Timed transmission networks

  • ●
  • ●
  • ●
  • ●
  • ≤ 2002

2003 2004 2005 2006 2007 2008 2009 ≥ 2010 Time of transmission

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 15/18

slide-54
SLIDE 54

Transmission between infection-routes

1995 2000 2005 2010 10 20 30 40 50 60 IDU−IDU

Number of transmissions

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 16/18

slide-55
SLIDE 55

Transmission between infection-routes

1995 2000 2005 2010 10 20 30 40 50 60 IDU−IDU

Number of transmissions

1995 2000 2005 2010 10 20 30 40 50 60 HET-IDU

Time of transmission Number of transmissions

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 16/18

slide-56
SLIDE 56

Transmission between infection-routes

1995 2000 2005 2010 10 20 30 40 50 60 IDU−IDU

Number of transmissions

1995 2000 2005 2010 10 20 30 40 50 60 MSM−MSM 1995 2000 2005 2010 10 20 30 40 50 60 HET-HET 1995 2000 2005 2010 10 20 30 40 50 60 HET-IDU

Time of transmission Number of transmissions

1995 2000 2005 2010 10 20 30 40 50 60 IDU−MSM

Time of transmission

1995 2000 2005 2010 10 20 30 40 50 60 MSM-HET

Time of transmission

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 16/18

slide-57
SLIDE 57

Summary

Genetic data was used to infer transmission networks

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 17/18

slide-58
SLIDE 58

Summary

Genetic data was used to infer transmission networks Threshold-based network suggests low cross-country transmission

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 17/18

slide-59
SLIDE 59

Summary

Genetic data was used to infer transmission networks Threshold-based network suggests low cross-country transmission Threshold-free approach highlights fragility of large networks

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 17/18

slide-60
SLIDE 60

Summary

Genetic data was used to infer transmission networks Threshold-based network suggests low cross-country transmission Threshold-free approach highlights fragility of large networks Incorporating temporal information suggests that IDU transmit to HET

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 17/18

slide-61
SLIDE 61

Acknowledgements

Thomas Lengauer Glenn Lawyer Mathieu Flinders Nico Pfeifer Tomas Bastys Rolf Kaiser Valeria Ghisetti Maurizio Zazzi Francesca Incardona Anne-Mieke Vandamme

euresist

Phylogeny-based HIV transmission networks Prabhav Kalaghatgi 18/18