Mark Woolhouse University of Edinburgh, UK

Home , Colorado tick fever, Marburg virus disease

WHAT WILL CAUSE THE NEXT PANDEMIC?

With thanks to: Feifei Zhang, Lu Lu, Liam Brierley, Alistair Morrison, Tara Wagner-Gamble and many others WHAT WILL CAUSE THE NEXT PANDEMIC?

The 2018 list of disease priorities needing urgent R&D attention comprises:

Crimean-Congo haemorrhagic fever (CCHF) Ebola virus disease and Marburg virus disease Lassa fever MERS-CoV and SARS Nipah and henipaviral diseases Rift Valley fever (RVF) The possibility that a serious Zika international epidemic could be Disease X caused by a pathogen currently unknown to cause human disease

Candidates included “highly pathogenic coronaviral diseases other than MERS and SARS”

Mark Woolhouse, University of Edinburgh, September 2020 WHAT WILL CAUSE THE NEXT PANDEMIC?

• Virus discovery • Known viruses • Disease X • Virus diversity • SARS-CoV-2 WHAT WILL CAUSE THE NEXT PANDEMIC?

• Virus discovery • Known viruses • Disease X • Virus diversity • SARS-CoV-2 HUMAN PATHOGEN SURVEY

• Systematic review of primary literature • Formal methodology → Taylor et al. (2001) Phil. Trans. B

• 70% EIDs are zoonotic • 70% EIDs caused by RNA viruses

 214 recognised RNA virus species (to 2015)  55 genera  21 families (+1 unassigned genus)

Mark Woolhouse, University of Edinburgh, September 2020 RNA VIRUS DISCOVERY

Mark Woolhouse, University of Edinburgh, September 2020 Woolhouse & Brierley (2018) Scientific Data PREDICTING WHERE

Risk maps created using a boosted regression tree model with 33 predictors. Models trained on 223 discoveries worldwide and 83 in China.

Top predictors: GDP/GDP growth > urbanisation > climate > biodiversity.

Discoveries since 2015 indicated (black triangles) and location of Wuhan shown.

Mark Woolhouse, University of Edinburgh, September 2020 Zhang et al. (under revision) PLoS Pathogens NATURAL HISTORY WITH NUMBERS

• 188/214 (88%) known human RNA viruses naturally infect other mammals • Most of the 26 human-specific RNA viruses have close relatives that infect other mammals [except: rubella, hepatitis delta] • Only 38/214 (18%) infect non-mammals [= birds (37) ± reptiles (7) ± fish (1?)]

• 55/74 (74%) mammal RNA virus genera include human viruses • 21/23 mammal RNA virus families include human viruses [except: Arteriviridae, Nodaviridae]

• “Majority of human viruses… are the product of host jumping” Kitchen et al. (2011) PNAS • Easier to switch host species than alter tissue tropism or transmission route

 Human infectivity evolves very easily within the mammal RNA viruses, less easily from birds and never(?) from anything else

Mark Woolhouse, University of Edinburgh, September 2020 WHAT WILL CAUSE THE NEXT PANDEMIC?

• Virus discovery • Known viruses • Disease X • Virus diversity • SARS-CoV-2 THE RNA VIRUS PYRAMID

LEVEL 4 N = 60 EPIDEMIC SPREAD

R0 > 1 LEVEL 3 N = 31 TRANSMISSION

LEVEL 2 N = 123 INFECTION

LEVEL 1 N = ??? EXPOSURE

Mark Woolhouse, University of Edinburgh, September 2020 updated from Woolhouse et al. (2014) in One Health CHANGING LEVELS

LEVEL 1 → LEVEL 2/3/4 LEVEL 4 • SARS EPIDEMIC SPREAD • MERS • Lujo • … many others LEVEL 3 TRANSMISSION LEVEL 3 → LEVEL 4 • Zaire ebolavirus LEVEL 2 • Chikungunya • Zika INFECTION

LEVEL 2 → LEVEL 3/4 LEVEL 1 • None? EXPOSURE

Mark Woolhouse, University of Edinburgh, September 2020 Woolhouse et al. (2016) Emerg. Infect. Dis. AN EXAMPLE: ARBOVIRUSES

• 91 human RNA virus species from 8 families are transmitted by vectors • All 19 level 3/4 RNA arboviruses are carried by anthropophilic vectors • Main anthropophilic vector species are from 5 dipteran genera: − Aedes spp. − Anopheles spp. − Culex spp. − Culicoides spp. − Phlebotomus spp. • All four Level 4 arboviruses [YFV, DENV, CHIK + ZIKA] are carried by Aedes spp. • There are no anthropophilic ticks: so no Level 4 tick-borne viruses

Mark Woolhouse, University of Edinburgh, September 2020 Woolhouse et al. (2016) Emerg. Infect. Dis. L3 RNA VIRUS SPECIES

Arenaviruses Coronaviruses Reoviruses (Dandenong*) Middle East respiratory syndrome (Colorado tick fever) Guanarito Nelson Bay Junin Filoviruses Rotavirus H Lassa Bundibugyo ebola Lujo Lake Victoria marburg Rhabdoviruses (Lymphocytic choriomeningitis) Sudan ebola Bas-congo* Machupo Zaire ebola (Rabies) Sabia Flaviviruses Togaviruses Bunyaviruses (Japanese encephalitis) Barmah forest Andes (Usutu) Chikungunya Bwamba (West Nile) O’nyong-nyong Crimean-Congo haemorrhagic fever Zika Ross river Oropouche Semliki forest Rift Valley Paramyxoviruses Venezuelan equine encephalitis Severe fever with thrombocytopenia Nipah syndrome

N = 3528 - includes 7 viruses known only through iatrogenic and/or vertical routes (parentheses)(parentheses) 8 with outbreaks >100 cases (bold)

*not ICTV recognised

Mark Woolhouse, University of Edinburgh, September 2020 adapted from Woolhouse et al. (2016) Emerg. Infect. Dis. L3 RNA VIRUS SPECIES

Arenaviruses Coronaviruses Reoviruses Middle East respiratory syndrome Guanarito Nelson Bay Junin Filoviruses Rotavirus H Lassa Bundibugyo ebola Lujo Lake Victoria marburg Rhabdoviruses Sudan ebola Bas-congo* Machupo Zaire ebola Sabia Flaviviruses Togaviruses Bunyaviruses Barmah forest Andes Chikungunya Bwamba O’nyong-nyong Crimean-Congo haemorrhagic fever Zika Ross river Oropouche Semliki forest Rift Valley Paramyxoviruses Venezuelan equine encephalitis Severe fever with thrombocytopenia Nipah syndrome

N = 282535 - includes 7 viruses known only through iatrogenic and/or vertical routes (parentheses)(parentheses) - 855 withwith outbreaksoutbreaks >100>100 casescases (bold)(bold) 3 already transitioned to Level 4 [EBOV, ZIKA, CHIK] *not ICTV recognised

Mark Woolhouse, University of Edinburgh, September 2020 adapted from Woolhouse et al. (2016) Emerg. Infect. Dis.

OUTBREAK DYNAMICS → MANY →

P(x)=Γ(x-½)/√πΓ(x)

FEW → FEW SMALL → → BIG

Mark Woolhouse, University of Edinburgh, September 2020 Woolhouse et al. (2016) Emerg. Inf. Dis. WHAT WILL CAUSE THE NEXT PANDEMIC?

• Virus discovery • Known viruses • Disease X • Virus diversity • SARS-CoV-2 PHYLOGENETICS OF DISEASE X

1764 sequences 39 RNA virus genera Levels 1, 2, 3/4 Mammal or bird host

Phylogenetic reconstruction of discrete state transitions using Bayesian MCMC / ML tree reconstructions and discrete trait / parsimony state reconstructions

Mark Woolhouse, University of Edinburgh, September 2020 Lu et al. (2019) doi.org/10.1101/771394 bioRxiv preprint PHYLOGENETICS OF DISEASE X Majority of new human viruses with epidemic potential are related to but not directly descended Betacoronavirus genus from other viruses that are transmissible in human (cf. Lyssavirus, Orthohantavirus) populations

57 human transmissible lineages (infective but not transmissible similar!)

Human NL63 dated at 1921 Lu et al. (2019) – Al-Khannaq et al. (2016)

Mark Woolhouse, University of Edinburgh, September 2020 WHAT WILL CAUSE THE NEXT PANDEMIC?

• Virus discovery • Known viruses • Disease X • Virus diversity • SARS-CoV-2 LIMITS TO RNA VIRUS DIVERSITY

• >200 human RNA virus species known • Have been recognising new species (mean = 2.8 per year since 1955); still discovering (but not yet ratifying) putative new ‘species’

M.E.J. Woolhouse, University of Edinburgh, November 2016 updated from Woolhouse et al. (2012) Phil. Trans. B LIMITS TO RNA VIRUS DIVERSITY

• >200 human RNA virus species known • Have been recognising new species (mean = 2.8 per year since 1955); still discovering (but not yet ratifying) putative new ‘species’

• Not finding new families, since a human picobirnavirus in 1988 • Marked slowdown in rate of discovery of new genera

M.E.J. Woolhouse, University of Edinburgh, November 2016 updated from Woolhouse et al. (2012) Phil. Trans. B LIMITS TO RNA VIRUS DIVERSITY

• Estimate of species pool Mora et al. (2011) PLoS Biol.

• hyperexponential model: log(logXr+1) = 2*log(logXr) - log(logXr-1)

No. genera per family No. species per genus

» Species pool = 275 (±10) [jack-knife 95% CIs] » 75% species already known • Consistent with extrapolations of discovery curve → up to 84% • cf. “unbounded” estimates, e.g. Woolhouse et al. (2008) Proc. B

M.E.J. Woolhouse, University of Edinburgh, November 2016 updated from Woolhouse et al. (2013) Future Virol. DIVERSITY OF MAMMAL VIRUSES

• <300 from extrapolating discovery curves Woolhouse et al. (2013) Future Virol. • >5000 species of mammal, 10 each → 50,000 Morse (1993) Emerging Viruses • Extrapolating bat/rodent surveys to all mammal and some bird viruses → 1.7M (SE 0.7-2.6M) Carroll et al. (2018) Science

? Do most mammals have any unique viruses at all? Critical community size

• Humans = 30% global land zoomass • Livestock = 67% • Wildlife = 3% Smil (2012) Harvesting the Biosphere

Mark Woolhouse, University of Edinburgh, September 2020 WT-VIZIONS Wellcome Trust-Viet Nam Initiative on Zoonotic Infections

Several thousand enteric, respiratory and CNS samples from hospitalised patients, high risk cohorts and animals UNUSUAL PATHOGENS

• Novel cyclovirus (CyCV-VN) in CSF from hospital patients • Novel porcine-like rotavirus (G26P[19]) in paediatric diarrhoea cases • First cases of Trypanosoma evansi infection in SE Asia • First human husavirus infections outside Europe • Novel kobuviruses in bats • Novel hunniviruses in rodents • Novel Bartonella spp in bats

M.E.J. Woolhouse, University of Edinburgh, February 2017 WHAT HAVE WE LEARNT?

• Exposure is very common; spillover is very rare • Pathogen discovery is not a systematic process • (Multi-faceted) surveillance is the first line of defence against emerging pathogens • “That district produces the greatest variety which is the most examined” – Gilbert White (1789)

Woolhouse et al. (2019) Scientific Data

• Should there be a Global Virome project? DIVERSITY OF MAMMAL VIRUSES

Mark Woolhouse, University of Edinburgh, September 2020 (PREDICTIVE) GENOMIC SURVEILLANCE

• Sequence data easy to obtain

• Inferring phenotype from genotype not at all easy to do

Babayan et al. (2018) Science

• Multiple ML algorithms on 500 virus genomes, 4000 genome traits • Up to 91% accuracy predicting vector type • Up to 72% accuracy predicting host type

Mark Woolhouse, University of Edinburgh, September 2020 (PREDICTIVE) GENOMIC SURVEILLANCE

• Sequence data easy to obtain

• Inferring phenotype from genotype not at all easy to do

• Key trait is cell receptor usage

• Know 78 different receptors from 94 human- infective viruses (across 19 families)

• Receptor → - Infectivity - Tissue tropism → - Transmissibility - Pathogenicity

• Large data sets on protein-protein interactions Mark Woolhouse, University of Edinburgh, September 2020 Woolhouse & Ashworth (2017) Biochemist PREDICTING VIRULENCE

• 58 ‘severe’ (27%) • 64 known to cause fatalities

• Machine learning: random forest, large number of independent models

• Allows many predictors/ combinations for given 89.4% viruses classified correctly dataset size (214 viruses) (null: 74.2%)

• Model predicts severity rating for each individual RNA virus

Mark Woolhouse, University of Edinburgh, September 2020 Brierley et al. (2020) PLoS Biol. WHAT WILL CAUSE THE NEXT PANDEMIC?

• Virus discovery • Known viruses • Disease X • Virus diversity • SARS-CoV-2 NOVEL CORONAVIRUS 2019-nCoV

• 08/12/19 - First patient in Wuhan, China • 02/01/20 - 41 patients, many with links to a market • 09/01/20 - first death • 09/01/20 - new coronavirus isolated • 10/01/20 - genome sequence published  most closely related to a bat coronavirus

• 29/01/20 – 7711 confirmed cases – 12167 suspected cases – 170 deaths (c. 1-2% cases) – 21 countries with confirmed cases

Mark Woolhouse, University of Edinburgh, September 2020 SOURCE Rhinolophus ferrumequinum ?

Mark Woolhouse, University of Edinburgh, September 2020 Wagner-Gamble (unpublished) WHERE AND WHY?

FMDV: Surrey, UK, 2007 SARS-CoV-2: Wuhan, China, 2020

Mark Woolhouse, University of Edinburgh, September 2020 ACKNOWLEDGEMENTS

Kyle Adair, Rustom Antia (EMORY), Jordan Ashworth, Melina Beykou, Alex Bhattacharya, Carlijn Bogaardt, John Brownstein (HARVARD), Liam Brierley, Liam Carr, Margo Chase-Topping, Sarah Cleaveland (GLASGOW), Peter Daszak (EHA), Chris Dye (WHO), Ben Evans, Eric Fèvre, Eleanor Gaunt, Sonya Gowtage-Sequeira, Dan Haydon (GLASGOW), Richard Howey, Zoe Hudson, Paul Kellam (SANGER), Sophia Latham, Lu Lu, Samantha Lycett, Chris McCaffery, David McCulloch, Nick Montgomery, Cristina Morena, Alistair Morrison, Conor O’Halloran, Kevin Olival (EHA), Amy Pedersen, Andrew Rambaut, Liam Reilly, Gail Robertson, Julie Robertson, Nora Schmit, Fiona Scott, Peter Simmonds, Donald Smith, Claire Taylor, Louise Taylor, Melissa Taylor, Tara Wagner-Gamble,Alex Wallace, Nicola Wardrop, Catriona Waugh, Nathan Wolfe (METABIOTA), Feifei Zhang

+ Foresight and IOM/NAS committees

Credit: T. Lembo FUNDING: Wellcome Trust, BBSRC, DEFRA/SFC, NIHR, USAID, Darwin Trust Mark Woolhouse, University of Edinburgh, September 2020