Reconstructing the social network of viruses in wild ducks

55
V Reconstructing the social network of viruses in wild ducks Eric J. Ma, Runstadler Lab BATS 2013 1

description

I describe the use of network modelling to answer questions about the ecology of the influenza virus in wild birds.

Transcript of Reconstructing the social network of viruses in wild ducks

Page 1: Reconstructing the social network of viruses in wild ducks

VReconstructing the social network

of viruses in wild ducks

Eric J. Ma, Runstadler Lab BATS 2013

!1

Page 2: Reconstructing the social network of viruses in wild ducks

What’s the deal with influenza viruses in ducks?

How do we reconstruct and learn from a social network of viruses?

What’s up next?

Outline

!2

Page 3: Reconstructing the social network of viruses in wild ducks

PB2 2.4 kb1

PB1 2.4 kb2

PA 2.2 kb3

HA 1.8 kb4

NP 1.6 kb5

NA 1.5 kb6

M1/M2 1.0 kb7

NS1/NS2 0.9 kb8

The flu virus has 8 (-) strand RNA segments in its genome.

!3

Page 4: Reconstructing the social network of viruses in wild ducks

Because of its segmented nature, the flu virus is capable of genetic reassortment.

!4

Page 5: Reconstructing the social network of viruses in wild ducks

!5

Brooks Range

Alaska Range

Minto FlatsFairbanks

Yukon Delta

Page 6: Reconstructing the social network of viruses in wild ducks

!6

Page 7: Reconstructing the social network of viruses in wild ducks

Minto Flats, Alaska: Breeding ground for ducks along the Pacific flyways.

!7

Page 8: Reconstructing the social network of viruses in wild ducks

!8

DONALD

Page 9: Reconstructing the social network of viruses in wild ducks

Duck flow = Virus flow

!9

Page 10: Reconstructing the social network of viruses in wild ducks

!10

Page 11: Reconstructing the social network of viruses in wild ducks

!11

Minto Flats

Page 12: Reconstructing the social network of viruses in wild ducks

!12

Page 13: Reconstructing the social network of viruses in wild ducks

15,083

Numbers

17520

2

ducks sampled

duck species in total

viruses isolated

years of sampling

!13

Page 14: Reconstructing the social network of viruses in wild ducks

But what is the relationship between

these viruses?

!14

Page 15: Reconstructing the social network of viruses in wild ducks

!15

VSign Up!

It’s free and always will be.

Page 16: Reconstructing the social network of viruses in wild ducks

What’s the deal with influenza viruses in ducks?

How do we reconstruct and learn from a social network of viruses?

What’s up next?

Outline

!16

Page 17: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map= 1 virus= different subtypes

!17

Page 18: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map= 1 virus= different subtypes

!18

Cluster each segment by genetic similarity (PWI)

Page 19: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map= 1 virus= different subtypes

!19

Different clustering patterns for each segment

Page 20: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map= 1 virus= different subtypes

!20

Threshold = minimum in-cluster PWI

Page 21: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map= 1 virus= different subtypes

!21

Estimate transmission transmissions

segment-by-segment

Page 22: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map= 1 virus= different subtypes

!22

Segment 1

Page 23: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map= 1 virus= different subtypes

!23

Segment 2

Page 24: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map= 1 virus= different subtypes

!24

Segment 3

Page 25: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map= 1 virus= different subtypes

!25

Segment 4

Page 26: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map= 1 virus= different subtypes

!26

Segment 5

Page 27: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map= 1 virus= different subtypes

!27

Segment 6

Page 28: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map= 1 virus= different subtypes

!28

Segment 7

Page 29: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map= 1 virus= different subtypes

!29

Segment 8

Page 30: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map= 1 virus= different subtypes

!30

All hypothesized transmissions

8

6

4

Page 31: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map= 1 virus= different subtypes

!31

Only full transmissions

8

Page 32: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map= 1 virus= different subtypes

!32

Page 33: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

Time

Order the isolates roughly in time.

= 1 virus= different subtypes

!33

Page 34: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

Plot out the full transmissions

Time

= 1 virus= different subtypes

!34

Page 35: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

In the Minto flats dataset, 460 out of 520 isolates were involved in 1096 full transmissions.

Time

= 1 virus= different subtypes

!35

Page 36: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

Of 1096 transmissions, 974 were estimated as direct transmissions.

Time

= 1 virus= different subtypes

!36

Page 37: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

Of 1096 transmissions, 49 were estimated as transmissions due to environmental persistence.

Time

= 1 virus= different subtypes

!37

Page 38: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

Of 1096 transmissions, 73 were due to multiple isolations from the same bird.

Time

= 1 virus= different subtypes

!38

Page 39: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

Of 1096 transmissions, 73 were due to multiple isolations from the same bird.

Time

= 1 virus= different subtypes

!39

Page 40: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

What if we coloured the nodes by host species?

Time

= 1 virus= different subtypes

!40

Page 41: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

What if we coloured isolates by host species?

Time

= 1 virus= different hosts

!41

Page 42: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

We can estimate inter- and intra-species transmissions

Time

= 1 virus

!42

= different hosts inter intra

Page 43: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

Time

= 1 virus

!43

= different hosts

379 inter-species transmissions644 intra-species transmissions

inter intra

Page 44: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

Time

= 1 virus= different subtypes

!44

Let’s colour isolates by subtype again.

inter intra

Page 45: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

In the Minto flats dataset, we have 92 subnetworks of full transmissions

Time

= 1 virus= different subtypes

!45

inter intra

Page 46: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

Were there any reassortant viruses?

Time

= 1 virus= different subtypes

!46

inter intra reassort

Page 47: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

In Minto Flats, we have putatively identified 52 reassortant viruses.

Time

= 1 virus= different subtypes

!47

inter intra reassort

Page 48: Reconstructing the social network of viruses in wild ducks

Numbers520 viruses isolated

1096 full transmissions between 460 viruses

974 direct transmissions

49 environmental persisters

52 putative reassortants

!48

Page 49: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

E.g. 1: Which viruses were really able to cross host species?

Time

= 1 virus= different subtypes

!49

inter intra reassort

Page 50: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

E.g. 2: How many viruses carry over to the next year?

Time

= 1 virus= different subtypes

!50

2009 2010

inter intra reassort

Page 51: Reconstructing the social network of viruses in wild ducks

Segment Transmission Map

E.g. 3: What factors allow cross-year viruses to persist?

Time

= 1 virus= different subtypes

!51

2009 2010

inter intra reassort

Page 52: Reconstructing the social network of viruses in wild ducks

What’s the deal with influenza viruses in ducks?

How do we reconstruct and learn from a social network of viruses?

What’s up next?

Outline

!52

Page 53: Reconstructing the social network of viruses in wild ducks

Q• How do host species barriers shape viral evolution?

• What is the relationship between Eurasian and N. American lineages of influenza?

• Is there a molecular basis for environmental persistence?

Page 54: Reconstructing the social network of viruses in wild ducks

d3.js-based, web-ready !interactive data explorer

(prototype ready)

because some questions don’t show up until you see the data

Page 55: Reconstructing the social network of viruses in wild ducks

Acknowledgments

!55

Jonathan Runstadler

The Awesome

Nichola Hill Brandt Meixell

who are amazing mentors & collaborators