Skip to main content

Notice: this Wiki will be going read only early in 2024 and edits will no longer be possible. Please see: https://gitlab.eclipse.org/eclipsefdn/helpdesk/-/wikis/Wiki-shutdown-plan for the plan.

Jump to: navigation, search

Difference between revisions of "Community Ebola Modeling Phone Call"

(November 19, 2014 Call)
(Agenda)
Line 34: Line 34:
 
# News
 
# News
 
# Items from participants
 
# Items from participants
# 20 minute deep dive topic: '''Impact of Exposed compartment on modeling disease dynamics'''
+
# 20 minute deep dive topic: Leah Shaw, College of William and Mary '''Impact of Exposed compartment on modeling disease dynamics'''
 
# Planning Next week's agenda
 
# Planning Next week's agenda
 
## Who would like to moderate in two weeks?
 
## Who would like to moderate in two weeks?

Revision as of 18:32, 12 November 2014

back to STEM Contents Page


Ebola.jpg


The STEM Community is holding a weekly phone call for open discussion on Ebola Modeling. Researchers studying Ebola Epidemiology, Modeling, and working on Ebola Response, are all invited. Our goal is to accelerate research by helping members of the scientific community interact, share data, questions, and ideas with each other, and to connect researchers with operational people. It is not necessary or required to be a user or contributor to STEM. All discussion should be open and non-confidential.

The Ebola community call is scheduled to take place most every Weds at 11AM Pacific Daylight time (2PM Eastern Time)

Join the Community Call

For more information, to add to the agenda, or if you wish to join, please send mailto:judyvdouglas@verizon.net

Participants

List of all current and previous Ebola Call Participants

November 19, 2014 Call

Phone call will begin at 2pm Eastern Standard Time (11AM Pacific Standard Time)

Agenda

for November 19

Moderator: Leah Shaw, College of William and Mary

  1. Welcome and Introductions
  2. News
  3. Items from participants
  4. 20 minute deep dive topic: Leah Shaw, College of William and Mary Impact of Exposed compartment on modeling disease dynamics
  5. Planning Next week's agenda
    1. Who would like to moderate in two weeks?
    2. What will be the deep dive topic?
    3. Please send short agenda items to jhkauf@us.ibm.com by Monday
    4. Please suggest other themes and guests for presentation/discussion
    5. Future Calls:
      1. Nov. 26, THANKSGIVING HOLIDAY
      2. Dec. 3, tbd
      3. Dec. 10, tbd
      4. Dec. 17, Mehmet H Gunes, title pending

November 12, 2014 Call

Phone call will begin at 2pm Eastern Standard Time (11AM Pacific Standard Time)

Agenda

for November 12

Moderator: Dr. Melissa Cefkin IBM Research, Almaden

  1. Welcome and Introductions
  2. News
  3. Items from participants
  4. 20 minute deep dive topic: Some Insights from Anthropologists on Ebola
  5. Planning Next week's agenda
    1. Who would like to moderate in two weeks?
    2. What will be the deep dive topic?
    3. Please send short agenda items to jhkauf@us.ibm.com by Monday
    4. Please suggest other themes and guests for presentation/discussion
    5. Next week's deep dive: Dr. Melissa Cefkin IBM Almaden. Topic "Some Insights from Anthropologists on Ebola"
    6. Future Calls:
      1. Nov. 19, moderator: Leah Shaw, College of William and Mary subject tbd
      2. Nov. 26, THANKSGIVING HOLIDAY
      3. Dec. 3, tbd
      4. Dec. 10, tbd
      5. Dec. 17, Mehmet H Gunes, title pending

Minutes

Attendees
  1. Sheldon Jacobsen, Dept of Computer Science, University of Illinois, Urbana
  2. James Kaufman
  3. Pat Selinger
  4. Melissa Cefkin
  5. Mary Helander December 3rd
  6. Richard Stokvis, Chief Medical Officer of Cures United, the Netherlands
  7. Simone Bianco
  8. Stefan Edlund
  9. Leah Shaw
  10. Nic Geard University of Melbourne
  11. Alexander J. Jones
  12. Michael Washington from CDC modeling Unit
  13. Mehmet Gunes
  14. Caitlin Rivers
  15. Bill Tetzlov
Discussion

Alex, total is 13594 confirmed cases. May underestimate by factor of 3.

5160 - 5400 Deaths Nurse died in Mali Imam died in Mali Africa Cup cancelled U.S. Nurses (Kaiser SF) on strike over Ebola measures and training Article on poverty and transmission, plan to model based on lights at night

Pat Syndemics references could motivate new models.

Caitlin WHO just posted new data portral

Alex. question. Early in the outbreak we were able to rely on underreporting as constant rate or value. But it seems to be changing. Can we look at data to estimate changing rate of underreporting

Deep Dive talk

Alex - JFK hospital referred to locally as "Just for Killing" even before Ebola Richard: comment Half the people die in the modeling, what could be done when you treat people properly - how would it change the outcome and develop trust in the treatment centers 400 nurses and doctors died because they did not have the required protected gear.

JK Q:? how does treatment change mortality rate Richard: Patient in Hamburg recovered. Patient in NY treatment not published yet successful treatments only partially published Is consensus among infectious disease experts that there is chance of recovery of 95% with proper treatment Statins and other drugs are thought to be quite effective. It's not just treating the virus but also treating the patient

Alex local practices that may even emerge from "Myths" sometimes lead to quarantine (not based on germ theory but still effective)

Bill T. Looked population statistics on health workers. modeling running out of health workers. We need numbers on health workers

discussion between Alex and Richard on Melatonin. Richard indicates there may be >50 compounds that help. With proper treatment people survive. The doctors mix and match. A lot of things seem to work. We need to see more published protocols. Fujifilm offered compound T-705 (Avagan registered in Japan for Flu, stockpiled for flu) WHO is trying to decide if when we should do a clinical trial - it's sitting idle. With > 50% mortality don't need placebo controlled trials, just do something. In Africa the protective gear equipment is like "cardboard" compared to gear in the US (not possible to remove it without touching skin). PH affects weather the chlorine works. 400 workers dies - most of them are locals. Treatment and equipment is not based on science.

10% of Ebola patients ALSO have active malaria. People with Malaria went for treatment and contracted Ebola in the clinic.

Items from Participants

November 5, 2014 Call

Phone call will begin at 2pm Eastern Standard Time (11AM Pacific Standard Time)

Agenda

for November 5

Moderator: Dr. Emma McBryde University of Melbourne

  1. Welcome and Introductions
  2. News
  3. Items from participants
  4. 20 minute deep dive topic: Ebola Infectivity over time Presenter: Dr. Emma McBryde. Please click HERE for the slides
  5. Planning Next week's agenda
    1. Who would like to moderate in two weeks?
    2. What will be the deep dive topic?
    3. Please send short agenda items to Judy by Monday
    4. Please suggest other themes and guests for presentation/discussion
    5. Next week's deep dive: Dr. Melissa Cefkin IBM Almaden. Topic "Some Insights from Anthropologists on Ebola"
    6. Next week's moderator: Melissa Cefkin IBM Almaden

Minutes

Attendees
  1. James Kaufman, IBM Almaden
  2. B.Tetzlaff, IBM Watson
  3. S.Bianco, IBM Almaden
  4. J. Douglas, IBM Almaden
  5. Jodie McVernon, University of Melbourne
  6. Kraig Butrum, Skoll Global
  7. Matt Davis, IBM Australia
  8. James Mccaw, University of Melbourne
  9. Stefan Edlund, IBM Almaden
  10. Ira Schwartz, US Naval Research Laboratory
  11. Emma McBryde, University of Melbourne
  12. Melissa Cefkin, IBM Almaden
  13. Kun Hu, IBM Almaden
  14. Alexander J. Jones, OperonLabs, NIH
  15. Leah Shaw, College of William and Mary
  16. Nic Geard, University of Melbourne
  17. Klimka Szwaykowska, US Naval Research Laboratory
  18. Pat Selinger, IBM Almaden
  19. Mehmet Gunes, University of Nevada, Reno
Discussion

News

B Tetzlaff: Read in Science health workers are still getting infected both in hazard suites but also in triage (without suits) Model does not segregate people who are with different risks. Can we extend model to distinguish clinical workers? Also people who are involved in burial, and friends and family in contact with sick family members outside of hospital

Emma: A number of people are thinking about how to model different risk groups. Alex: As model gets more complex have more granularity but sometimes data you feed the model is not available. Jamie: Agrees with Alex. Bill: Agrees....

Emma: All agree would be nice to have more model complexity but there are problems in doing that in particular finding appropriate parameters from the data and introducing unknown parameters. Word in newspapers is that things are slowing down in Liberia.

Alex: WHO gone from 3%/day increase to 1% but 2.5% factor under-reporting. Seems like a slowing in the rate of growth in Liberia Sierra Leone is accelerating. Difficult to separate out the noise. We really don't know and won't know for 1-2 months.

B Tetzlaff: Under-reporting is a guess and estimate is continuously changing.

Alex: If we take reports at face value there is a deceleration in Liberia and acceleration in Sierra Leon but we don't really know.

Emma: H Nishira did a model in Eurosurveilance looked at three countries as a system. R0 in Guinea<1 but for whole system R0 >1. This could happen again as various countries try to bring Ebola under control.

Alex: Rate accelerating in area around Freetown. Avg # daily cases is 6x higher than 2 months ago.

Deep Dive talk

Ebola Infectivity over time Presenter: Dr. Emma McBryde. Please click HERE for the slides

Infectiousness increases continuously over time from infectious onset to death slide 2 shows level in blood as fn of days after detection then drops off (for both fatal and nonfatal) Paper by Yemen and Galvani Anals of Internal Medicine. Oct 28th published online Example of a model that does microsimulations including infectivity by day Also looked at contacts drawn contract from a distribution. Assume same for everyone except during last stages of disease should we model change infection over time. If we want detail in interventions (eg training family for nursing at home) there is a delay between onset of fever and interventions - this should be captured in the model. showed how to extend the previous ebola model only additional parameters in this simple extension were transitions between infectious compartments. Three new transitions

Slide 7:

Q: Pat people die too early so peak happens too soon

E: That is correct. Slide 8 shows this

Patt: If we turn off hospitalization we would see this peak (breakdown in hospitalization)

Emma: Correct.

Alex: Useful to look at that.


The model results are very sensitive to assumptions around interventions on duration in the infectious compartment
Sensitivity extends to assumptions about shape of infectiousness and shape of interventions
These may need to be considered when thinking of interventions like home quarantine

Q from Simone: should we have fewer deaths from I1,I2 than I3,I4 how would that change the results ? Emma: Agree. Selecting from a parametric distribution of observed cases would give a more realistic model. The goal here was just to see if changing infectious changes the outcome of interventions. Lots of parameter changes would not matter much in this model but in reality might be very important. Starting from scratch want to use more realistic values in each compartment

Items from Participants

Item from Emma. Q for STEM people about issue of data sharing

Is creative commons ok for Eclipse.

Jamie: Probably it is. We just need to follow the submission process and Eclipse attorneys will review the license.

Jamie will update wiki on stochastic solvers

Leah Shaw will moderate in two weeks

October 29, 2014 Call

Phone call will begin at 2pm ET (11AM PDT)

Agenda

for October 29

Moderator: Simone Bianco, IBM Almaden Research Center

  1. Welcome and Introductions
  2. News
  3. Any follow-up items?
  4. Items from participants
  5. 20 minute deep dive topic: Modeling news on recent sensitivity analysis, SEIR++Parameters. Presenters: Kun Hu and Simone Bianco. Please click here for the slides [1]
  6. Next week's agenda
    1. Who would like to moderate in two weeks?
    2. What will be the deep dive topic?
    3. Please send short agenda items to Judy by Monday
    4. Please suggest other themes and guests for presentation/discussion
    5. Next week's deep dive: Dr. Emma McBryde, University of Melbourne. Topic to be announced
    6. Next week's moderator: Dr. Emma McBryde

Minutes

Attendees
  1. Simone Bianco (Moderator) IBM Almaden
  2. Dr Abdessamad Tridane for UAE University
  3. Ada Yan University of Melbourne
  4. Jodie McVernon, UoM
  5. Alexander J. Jones Operon Labs
  6. Lauren Barthel Operon Labs
  7. Ira B. Schwartz US Naval Research Laboratory
  8. Luis Mier-y-Teran John's Hopkins, NRL
  9. Klimka Szwaykowska US Naval Research Laboratory
  10. Emma McBryde University of Melbourne
  11. Bradford Green CDC Modeling Task force
  12. Mehmet Gunes, University of Nevada, Reno
  13. Nick Geard, University of Melbourne
  14. James Kaufman IBM Almaden
  15. Mary Roth IBM Almaden
  16. Melissa Cefkin IBM Almaden
  17. Kun Hu IBM Almaden
  18. Judy Douglas IBM Almaden
  19. Pat Selinger IBM Almaden
  20. Stefan Edlund IBM Almaden
  21. Bill Tetsloff IBM Watson
  22. Mary Helander IBM Watson
  23. Matthew Davis IBM Australia
  24. Rosalyn Hickson IBM Australia
Discussion

NEWS

Nick G: 2-year-old girl was sick on bus to Mali (now tracing 40+ contacts). She passed away.

Bill Tetsloff, NY, NY: People distrust information from CDC, etc.; are very afraid and need to be educated.

Alex (?): Lack of education in the media; cover is way to one side or the other. Becoming a political issue.

Judy: Woman taken to Maine will file suit tomorrow if they don't release her from quarantine.

Alex: Sent out public data sets on cell phone data. Most are private. Found one set with 146 people.


Deep Dive talk by Kun Hu and Simone Bianco

Kun and Simone presented STEM epidemiological model


Further Discussion

Emma: might find more uncertainty if use a different error function (binomial vs Poisson) Post mortem transmission rate is outside the literature range.

Kun: Yes, that's what our fit showed.

Alex: created a STEM model SEIR - model looks useful. NEJM paper has a weighted average for Ro from several models. comes up with 1.84. Q how Kun got Ro.

Kun Answer is: in the slides, expression for R0 + data in table.

Simone: C. Chavez, Vespignani and others have different values but all have R0<2.

Tridane: This model assumes everybody susceptible. Majority of ?happened at beginning of epidemic. Can you split S to those that work in Healthcare and those that do not. Many people do not go to hospital. Suggestion to split the susceptibles

Kun: Yes, one can split S into clinical workers and others. This could be a future extension to the model.

JK this would be particularly valuable if we could get statistics on clinical workers infected vs non-clinical. Otherwise it's easy to create the model but difficult to calibrate.

Emma: Epidemic peak time. Under any parameter that the epidemic had not peaked after 3 years

Kun: Not for these three countries. All the peaks times for three countries happen within 3 years.

Tridane: Do you plan to use model to examine efficiency of screening His group has model for vector born disease which shows difficulty of border screening

Kun: We've had this discussion during our call on Oct/15th. We've done some preliminary exploration and confirmed by participants from CDC during the call.

Alex great model but it is a limitation we don't have more data to seed the compartments especially burial rates

Simone burial rate is the most difficult to get. In SL if they suspect Ebola they have to report and then someone must come test body Very difficult to happen quickly. The delay is up to a week. Some people wait. Some people proceed with burial, some actually put bodies on street.

Items from Participants

Melissa will moderate and give deep dive in 2 weeks

Emma's topic for next week "has anybody mapped infectivity over time?"

Alex has paper about asymptomatic infectious over time

Tentative title "Ebola Infectivity over Time"

October 22, 2014 Call

Phone call will begin at 2pm ET (11AM PDT) to accommodate participants from Australia

Agenda

for October 22

Moderator: Ira B. Schwartz, US Naval Research Laboratory

  1. Welcome and Introductions
  2. News
  3. NIHR UK Follow-up: Asymptomatic carriers will pass through screening undetected
  4. Items from participants
  5. 20 minute deep dive topic: Adaptive human behavior to control epidemics Ira B. Schwartz, US Naval Research Laboratory. Please click here for the slides File:Ebola AN deep dive.pdf
  6. Next week's agenda
    1. Simone Bianco, IBM will moderate next week. Who would like to moderate in two weeks?
    2. Please send short agenda items to Judy by Monday
    3. Please suggest other themes and guests for presentation/discussion
    4. Next week's deep dive topic: Kun Hu & Simone Bianco model news on recent sensitivity analysis, SEIR++ Parameters
    5. Next week's moderator Simone Bianco
    6. In two weeks Dr. Emma McBryde, University of Melbourne will moderate and present Deep Dive next topic: "need title"

Minutes

Attendees
  1. Alexander J. Jones, OperonLabs
  2. Jamie Kaufman, IBM Research
  3. Ira Schwartz, US NRL
  4. Ada Yan, University of Melbourne
  5. Niina Haiminen, IBM Research
  6. Kun Hu, IBM Research
  7. Judy Douglas, IBM Research
  8. Matt Davis, IBM Research
  9. James McCaw, Melbourne School of Population and Global Health
  10. Leah Shaw, College of William and Mary
  11. Simone Bianco, IBM Research
  12. Pat Selinger, IBM Research
  13. Chang Chang, Peking University
  14. Melissa Cefkin, IBM Research
  15. Raul Andino, UC San Francisco
  16. Emma McBryde, University of Melbourne
  17. Nic Geard, University of Melbourne
Discussion

Jamie Kaufman: OperonLabs model contribution, to be contributed officially to Eclipse

Alex Jones, James McCaw, Jamie Kaufman: Need for an air travel model Adding escape rate with stability analysis of possible interest. There is a need to track people. US is enforcing travel restrictions to passengers from West African countries, to land in specific airports. Proposed joint collaboration to address air travel model.

Alex: Bats are a known reservoir. However, bats are not affected, just carriers, with mortality increasing upon transmission to human hosts. The reason why are bats not affected is still unknown. Bats are 30% of all mammals - why are they not affected by most virus? Bat interferon activates different genes. Interferon thought to be the link to an increased T cell response.

Time conflict with NEJM web update noted

Deep Dive talk by Ira B. Schwartz

Adaptive human behavior to control epidemics Ira B. Schwartz, US Naval Research Laboratory. Please click here for the slides File:Ebola AN deep dive.pdf


Discussion Notes

  • Emma McBryde: Hepatitis C network dynamics analysis shows that the most connected people can be found by following an edge.
  • Likely to be connected to people with high degree distribution so the infection quickly goes to a hub.
  • Hep C has been shown to spread among friends.
  • Find high degree nodes and use ring vaccination strategy.
  • Alex: Social distancing affects on Ebola (being afraid of your friend).
  • Ira: Leah Shaw and he looking at social distancing, but it's hard to get any kind of data.
  • Study on manipulation (put wash basin outside of men's room - people can see if you wash your hands or not). Use peer pressure to modify behavior.
  • Some behaviors early in the epidemic were counterproductive
  • Leah: Although there is a lack of data, modeling in her group shows that, if your social distancing increases as epidemic ramps up, and then goes back,

you can trigger oscillations.

  • James M.: Absolutely true. Shown for Influenza that social distancing can cause oscillatory dynamics (Australia, 2009 swine flu, other examples, Stephen Reilly paper
  • James will put references on the wiki
  • Emma: Seen in SARS but hard to resolve when simultaneous government intervention and decision making
  • Nic: Question on on clustering emerging in Ira's model: Does it break the network? At what scale?
  • Rewiring - can be reduction in efficacy if nearby people are also infectious (probably true)
  • Alex: Cell phone data sets hard to get. Suggestion to use 4SQ
  • Ira: In the first paper (Barabasi group), cell phone data was not quite anonymous
  • Want to get distribution function that describes peoples movements (??network topology??)
  • Alex: Nokia mobility data challenge data set is public. Will send link
  • Emma: global air travel. Group in Melbourne is quite interested. Would like to fill in regions for Australia. Offer to help refresh the data
  • Kun: several teams are deploying vaccines. Different vaccination strategies could me modeled and it might be worth discussing over.
  • Ira: Stratification of population - age dependence.
Items from Participants

Next Week Simone will moderate.

In two weeks Emma will moderate.

October 15, 2014 Call

Phone call will begin at 2pm ET (11AM PDT) to accommodate participants from Australia

Agenda

for October 15

Moderator: Alexander J. Jones, Operon Labs

  1. Welcome. Introducing
    1. Dr. Emma McBryde, Head of Mathematical Modeling, Burnet Institute, Univ of Melbourne, Victorian Infections Disease Service
    2. Prof. Jodie McVernon, Modelling and Simulation Group, Centre for Epidemiology and Biostatistics and Vaccine and Immunisation Research, Murdoch Children's Research Institute and Melbourne School of Population and Global Health
    3. Prof. James Mccaw, Modelling and Simulation Group, Centre for Epidemiology and Biostatistics and Vaccine and Immunisation Research, Murdoch Children's Research Institute and Melbourne School of Population and Global Health
  2. News
    1. Added link to stem-ebola summary on the home page
    2. Added Summary of Pub-Med news Items to our Ebola Reference Data page.
    3. Please copy (quantitative) new updates to the Ebola Reference Data. This way we can track the dates of key events (ie if we want to tally imported and secondary cases outside of W. Africa over time.
  3. Questions on expanding the community: Pro Med and other mechanisms (Alex, Jamie)
  4. Caitlin Update on VT hack-a-thon
  5. 20 minute deep dive topic: Ebola Deep-Dive Topic: Mutation and Fitness Landscape
    1. Ebola 2014 Mutation Rate: Comparison to previous Ebola outbreaks & Other Viruses
    2. Potential Impact of Ebola Mutations (tissue tropism, fitness landscape)
    3. Superinfection: Mathematical Properties / Evolutionary Dynamics
    4. Ebola Virulence vs Infectivity: Confounding Variables
    5. Recombination: Evidence for Horizontal Gene Transfer in Ebola
  6. Next week's agenda
    1. Ira B. Schwartz, NRL will moderate next week. Who would like to moderate in two weeks?
    2. Please send short agenda items to Judy by Monday
    3. Please suggest other themes and guests for presentation/discussion
    4. Next week's deep dive topic: Kun Hu & Simone Bianco model news on recent sensitivity analysis, SEIR++ Parameters
  7. Items from participants
    1. Question from NIHR UK

Minutes

Attendees
  1. Alexander J. Jones, Operon Labs, Moderator
  2. Jodie McVernon, Murdoch Children's Research Institute and Melbourne School of Population and Global Health
  3. James Kaufman, IBM Research
  4. Kun Hu, IBM Research
  5. Melissa Cefkin, IBM Research
  6. Judy Douglas, IBM Research
  7. Roslyn Hickson, IBM Research Australia
  8. Leah Shaw, William and Mary
  9. Stefan Edlund, IBM Research
  10. James Mccaw, Murdoch Children's Research Institute and Melbourne School of Population and Global Health
  11. Ira B. Schwartz, US Naval Research Laboratory
  12. Luis Mier, US Naval Research Laboratory
  13. Caitlin Rivers, Virginia Tech
  14. Raul Andino, UCSF
  15. Martin Meltzer, CDC Modeling Unit, Ebola Task Force
  16. Manoj Gambhir, CDC Modeling Unit, Ebola Task Force
  17. Thomas Gift, CDC Modeling Unit, Ebola Task Force
  18. Stuart Nichols, CDC Modeling Unit, Ebola Task Force
  19. Simone Bianco, IBM Research
Discussion

Question from NIHR in the U.K.. At what rate will (asymptomatic) individuals infected with Ebola pass undetected through airport screening at arrival airport if that screening involves only taking temperature?

IBM Research analysis: If we assume the passengers do not know they are infected, and if exponential growth continues, we estimate 80% will go undetected.
CDC Modeling Unit analysis: Based on the CDC model, if the passenger boards the plane in the asymptomatic state, 
about 90% will pass through the arrival airport undetected. At best 20% would be detected.

Caitlin Rivers gave a short talk with slides describing the recent File:VirginiaTechHackathonOct15.pdf held on Oct 15th.

Deep Dive talk by Alexander J. Jones

Oct 15th Deep Dive Discussion Slides up on the Operon site: http://www.operonlabs.com/?q=node/18

File:Deep Dive Oct 15 Operon Labs v1.pdf <--- * Download PDF of Operon 'Deep Dive' Slides

Discussion Notes

  • Discussed two definitions: Virulence vs infectivity
    • Virulence - mortality, morbidity
    • Infectivity - basic reproduction number. Inherent ability to spread
  • Impact of mutations so far: RNA virus - error prone polymerase. Churns out SNPs as well as insertions and deletions
  • Frame shift mutation might lead to unfit offspring that will die. Some may have better some worse fitness
  • Current outbreak, most common ancestor is 2007/2008 Congo outbreak
    • Absolute parent E. Zaiher is the 1976 concensus strain
    • Current virus 97% similar to concensus strain
    • ~400-500 mutations/substitutions = 3% difference
  • Fig 2b 3 clades small guinea clade. 3-4 subtypes. Feb-March is different from May/June. May be functionally the same but it is changing
  • Fig 3 shows 7 transcribed genes
  • 4 subclades of virus can be identified in different regions.
  • Don't know the impact on fitness
  • Virus mutation rate is accelerating
    • Rate is twice the mutation rate before this outbreak
    • This agrees complete with predictions by Bianco and Andino
  • Andino: The mutation rate is not uniform along the viral sequence, but depends on the type of mutation and position in the genome; Also, standard NGS may not be accurate enough to capture all the mutations; Finally, synonymous mutations may be important just as well as non-synonymous mutations, as they may have non-zero fitness effects and may contribute to the complex genetic landscape of the virus.
Items from Participants

Change of plans for next week's call. Ira Schwartz will moderate next week and give the deep dive

Simone Bianco will moderate in two weeks and give the Kun Hu/S. Bianco deep dive that week

So by convention going forward the Deep Dive speaker will be their own moderator (keeping their time to 20 minutes)

October 8, 2014 Call

Agenda

for October 8

  1. Welcome
  2. Vote on time change for call
  3. News
  4. New Eclipse tools
    1. Ebola community mailing list ( please sign up )
    2. Data page on wiki
    3. Literature references page
  5. 20 minute deep dive topic: Caitlin on Contact Tracing (lots of chatter from various sectors this past week)
  6. Next week's agenda
    1. Who would like to moderate?
    2. Please send short agenda items to Judy by Monday
    3. Please suggest next week's themes for presentation/discussion
    4. Next week's deep dive topic: Alex on SEIR Parameters (average or median regional values? values from a single paper? how to select for simulations?)
  7. Items from participants

Minutes

Attendees
  1. Caitlin Rivers, Virginia Tech Moderator
  2. L. Shaw, William and Mary
  3. Simone Bianco, IBM Research
  4. Kun Hu, IBM Research
  5. James Kaufman, IBM Research
  6. Judy Douglas, IBM Research
  7. Stefan Edlund, IBM Research
  8. Alexander J. Jones, Operon Labs
  9. Ira B. Schwartz, US Naval Research Laboratory
  10. Melissa Cefkin, IBM Research
  11. Mehmet Gunes, University of Nevada, Reno
  12. Pat Selinger, IBM Research
Discussion

Motion to hold the call one hour later for Australian Participants.

Motion passes without objection

Jamie: reviewed new community tools

Caitlin: Deep dive on contact tracing one of most basic public health interventions (interview patients to identify contacts) Contacts followed through incubation period In W Africa there are >20,000 active contact right now. Not going very well. G, SL public projects on following contacts they miss hundreds every day

Q from Simone: Once identifies as contact are they asked to quarantine themselves A no they are not isolated they go about their normal days until symptomatic

Q Alex: What is the maximum contact tracing can reduce an outbreak by?? A Varies by disease. If infectious during incubation it's not effective but with Ebola it is infective

Q is 50% reasonable A if worked perfectly it would be 100% effective. In practice it's 50-75% effective

Q can we use the contact tracing effectiveness data do detect secondary cases (ie from asymptomatic if there are any) A Ira: It's difficult due to uncertainty in the incubation period - was the secondary case asymptomatic or not.

Q Ira Do we know how tight communities are? DO they try to isolate communities from General population A Some natural isolation

Jamie: Firestone example

NPR
USA Today

Caitlin Two primary ways it can go wrong

  • contacts not seen (true now)
  • contacts lost to follow up (ie avoiding the tracing teams)

Montserrado county in Liberia has >200 lost to follow up

This is a major problem for control Call to modelers to think about this as a network problem or resource allocation problem

How can we be more efficient - place the monitors and bring the contacts to the modelers? Melissa: The network of people being traced. If we knew who are their family members (is that the network)? Caitlin: Could track them as well the contact teams should have that info

Ira: Watching videos. 1 contact tracing team for what seemed like 1/4 of the country. Has the situation improved? Caitlin: Not improved. LIberia does not have enough vehicles. No organization in how town is laid out (address associated with people not places)

Ira: Is there a model that would work in west africa that is not being tried? Local teams vs top down approach?

Jamie: Could we airdrop thermometers and send a text if you have a fever. IBM Kenya lab has cell phone reporting system... ie contact trace everyone.

Caitlin has DOD contacts asking for this. Jamie will connect to Kenya lab contacts

Done

Simone: Met with Raul Andino about the probability Ebola will become airborne. Prob is zero. Will it evolve so asymptomatic individuals shed the disease he said this is more likely. Any RNA based virus can evolve to have higher viral load without showing symptoms. We could implement Ira and Carlos' ideas to ask how would epidemic change if we get asymptomatic transmission. Alex: Highly probably but not enough cases yet (3000 people is not enough. with millions of people we might see case fatality rate go down but transmission could go up. There is a great book called Evolutionary dynamics (how virus explore). But depends on very large number of cases. Fatality might go down but not down that far)

Next Weeks Agenda Alex will moderate next week. Topic will be virology. Caitlin will also brief us on the hack-a-thon

Ira will moderate week after next will think about deep dive topic (and we can ask melissa)

Items from Participants

Items form Participants Kun would like to know about the hackathon at Virginia Tech Caitlin Today is the first day. Next three days are all out on it. Check back early next week.

Caitlin will invite people from DOD to this call. Ira: there are tri-service people that might be interested in joining (they are focused on pandemic modeling). Ira/Simone will exchange email viral evolution.

October 1, 2014 Call

Agenda

  1. Introductions
  2. Timing for Community Calls
  3. Purpose
    1. Not to push one model
    2. Not to advocate one tool
    3. Support Ebola response efforts
  4. Eclipse Community Tools
    1. This Call
    2. Newsgroup
    3. Mailing list instructions
  5. Overview of Ebola Model and four Ebola Scenarios uploaded to Eclipse
    1. Admin 0 three country model for West Africa
    2. Admin 2 three county models for West Africa
    3. All Africa Model
    4. The Global Model
    5. How to easily change from deterministic to stochastic
    6. Running STEM Headless on server
  6. Discussion on Literature models - please add references to this page
  7. Discussion on model parameters (latest wisdom, sensitivity analysis)
    1. Should we create a wiki page for ongoing discussion?
    2. Should we create a newsgroup topic for ongoing discussion?
  8. Next week's agenda
    1. Who would like to moderate ? We can rotate.
    2. Please send short agenda items to Judy by Monday
    3. Please suggest longer themes for presentation/discussion
  9. Items from participants

To receive agenda updates form our mailing list please:

Get an Eclipse ID: [2]
Subscribe to the mailing list [3]
See information on the STEM Community Join the STEM Community

Model documentation will be available on the wiki page Ebola Models

Minutes

Attendees
  1. James Kaufman, IBM Research
  2. Kun Hu, IBM Research
  3. Simone Bianco, IBM Research
  4. Judy Douglas, IBM Research
  5. Stefan Edlund, IBM Research
  6. Caitlin Rivers, Virginia Tech
  7. Sherry Towers, Arizona State University
  8. Bob Pinner, CDC
  9. Mehmet Gunes, University of Nevada, Reno
  10. Ira B. Schwartz, US Naval Research Laboratory
  11. Christian Althaus, ISPM, University of Bern
  12. Pat Selinger, IBM Research
  13. Vincent Ruslan, Operon Labs
  14. Carlos Castillo-Chavez, Arizona State University
  15. Melissa Cefkin, IBM Research
  16. Bryan Lewis, Virginia Tech
Discussion

Christian:

  • Science paper did not look at how mutations have changed properties of the virus
  • Population structure and control measures are quite different from countries
  • Parameters similar to previous outbreaks

Carlos:

  • This is a much bigger outbreak

Sherry:

  • The current Ebola outbreak seems to have relative low fatality rate compare to 90% in the record
  • Is asymptomatic transmission playing a role?

Bryan:

  • Is case mortality lower?

Sherry:

  • It seems to be lower.

Caitlin:

  • Up to 80% case fatality
  • Infectious period seems to be twice times than the previous outbreak.

Vincent:

  • What is the role of asymptomatic transmission. some numbers suggest something is different.

Jamie:

  • What is R0 for community, hospital, vs funeral transmission?

Sherry:

  • Suggested the transmission at funeral is 2-5 times higher.
  • Will provide some papers discussing different R0

Christian:

  • Really difficult to quantify different elements of transmission - restricted to total incidence data

Simone:

  • Conflicting data on under reported cases.

Sherry:

  • Who is collecting data?

Caitlin:

  • MOH of respective countries

Bryan Lewis:

  • Telecon with Neil Ferguson. Analysis of case listings.
  • The current endeavor is to explore these questions.
  • The data is partial

Christian:

  • STD transmission is probably minor

Everyone:

  • We need to get all the literature references in one place
Items from Participants

Caitlin:

  • Suggests we create a wiki page on Data
  • Ok to link to Caitlin's git hub and blog
Done: see new pages and please feel free to add content
Literature Ebola References 
Data Ebola Reference Data 

Sherry:

  • As a statistician, the work is data driven
  • We need better data to inform our models

Vincent:

  • Also concerned about asymptomatic transmission

Ira:

  • Interested in how people adapt their behavior in response to the epidemic outbreak
  • Also concerned about asymptomatic infectious classes. Can we back this out to predict asymptomatic?
  • Interested in agent-based model to study these type of questions

Christian:

  • Interest in opportunity for transmission in different small outbreak in Nigeria on how R0 (reproductive number) changes in a better urban setting where interventions were effective.
  • How does R0 depend on healthcare system of the country?

Pat:

  • Being in a community, how can we efficient and effective work together, what would help people respond more quickly?
  • Using a maillist does not address the issue when people want to share dataset.

Vincent:

  • Suggest to have a Ebola mailing
  • Parameters show this outbreak is going to be a 12-24 month long

Carlos:

  • Asymptomatic individuals: are they infectious or not? Very important in Influenza
  • Explore time dependent value of parameter
  • Critical to determine which of the additive factors contribute the most to R0 and Reff
  • Parameters change with different populations, different practices, different environment that may facilitate the transmission.
Follow up
  • Stefan Edlund created a new mail list stem-ebola@eclipse.org that will be active within 24 hours
  • Simone Bianco posted these minutes
  • James Kaufman created a new wiki page Ebola Reference Data

Back to the top