Please tweet - everything! - PowerPoint Presentation

Slide1

Please tweet - everything!#openashdb

@

danmaclean

- bioinformatics

@kamounlab – pathogenomics

Crowdsourcing for ash diebackSlide2

Crowdsourcing for ash dieback

Kentaro Yoshida

,

Diane Saunders, Sophien Kamoun and Dan MacLean

GMOD meeting 5.April.13Slide3

Ash

tree (

Fraxinus

Excelsior) Slide4

Yggdrasil in Norse mythology is a giant Ash.

"The Ash Yggdrasil" (1886) by Friedrich Wilhelm Heine.

Healing treePre-Christian: Pass a sick child through split tree: if it resealed the child would be cured. Strong Furniture

Withstand shocks Oars, cues, truncheons, hockey sticks etc

Central in Norse cosmologySlide5

Lesions and cankers on stems/

branches V

isible throughout the year

Leaves with brown leaf stalks

Throughout summer

Fruiting bodies on fallen

leaf stalks

Visible from spring

Ash diebackSlide6

Ash dieback symptoms

Photos: Iben M Thomsen

In DenmarkSlide7

Chalara fraxinea

Alias: Hymenoscyphus pseudoalbidusSlide8

Ash dieback disease – Chalara

fraxinea

2012Slide9

Ash dieback

http://

ashtag.orgSlide10

S

cience is too slow in emergencies

We have to wait for funding of relatively isolated groups on specific projects

Structure of science inhibits collaboration and sharing

Publication cycle bad for usSlide11

“many hands make light work”

Crowdsourced analyses, open access data

let the experts at the dataSlide12

Crowdsourced

analyses

“live peer review – the global on-line lab meeting”

Let the experts review the results as they appear – live filteringSlide13

Why crowdsourcing might help

>3000 people hospitalized

50 deaths in Germany

Outbreak tracked to Fenugreek seeds (used as a herb, spice or vegetable)

Scientific response

Dr

Loman

joined up sequences (@

pathogenomenick

)

24h

48h

72h

96h

120h

144h

168h

DNA-based diagnostics

Key findings identified:

How it kills

T

oxin genes

(Example) Applying

crowdsourcing to deadly diseases:

E. coli outbreak Germany 2011

github

:

ehec

-outbreak-crowdsourced

/ BGI-data-analysisSlide14

an initiative to fast-forward collaboration on

chalara dieback of ash

OpenAshDieBackhttp://

oadb.tsl.ac.ukSlide15

Data

Which license ?

NONE WHATSOEVER! NOT Fort Lauderdale, NOT Toronto. COMPLETELY OPEN ACCESS, PUBLIC DOMAIN! Slide16

github

version management and contribution tracking

p

ull

data

make change

push

back

The data and results themselves are actually hosted externally on the public website,

github

. Slide17

What the repo is -

Basically just as directory structure – semantically organized ‘github.com/ash-dieback-crowdsource/data’A fork of a generic repo for this stuff

‘github.com/danmaclean/crowdsrc’

you can start your own right nowSlide18
Slide19

Github accesses

Number of signups: 21 Directory size (not including reads): 4.32 GbNumber of commits: 103

Quite a large

labgroup So from nothing were generated a whole new research groupSlide20

All analyses contributed(what we learnt since December!)

is on the wiki and blogSlide21

a hub for analysis reports

Diane Saunders @ TSLhttp://oadb.tsl.ac.ukSlide22

Look for genes with similarity to known disease causing proteins

C.

fraxinea toxin (NLP1)Recognized a toxin based on its similarity to a common fungal toxin (toxic to plants)

C.

fraxinea

NLP1

Fungal NLP

I

dentical regions in blue

C.

fraxinea

NLP1

Fungal

NLP

toxic part of proteinSlide23
Slide24

Getting bioinformaticians is fine, want also to get bench biologists involved (these know all about pathogen!)

need new infrastructureSlide25

OADB

c

loud

t

ools

Data Store

Dedicated interim raw data storage

GitHub

assembly and annotation hosting (

bioinformaticians

)

Assembly and annotation web-tool (bench biologists)

Administrative middleware

Hub website and access point

?

G-

ny

-MOD - ‘Generic not-yet-a Model Organism Database’

Holds data

while model under construction

ftp-

oadb.tsl.ac.ukSlide26

gee

fu

portable feature

and assembly versioning database

RESTful

API – script access

Works well for small groups of biologists

Very small internal tool – not yet ready for primetime, but lightweight

github.com

/

danmaclean

Dan MacLeanSlide27

gee

fu - ‘

experiments’Slide28

gee

fu - ‘

tools’Slide29

gee

fu -

‘tools’Slide30

gee

fu - ‘tools

’Slide31

gee

fu browsingSlide32

Right now- we’re building this

But we need a good tool – WebAppollo??We ask you now to give us suggestions (we’re crowdsourcing you right now)We REALLY would like a better solution than “gee fu

”! Let us know! How can GMOD accommodate these needs!Slide33

http://oadb.tsl.ac.uk

How to get involved

g

o and get the data!

do your stuff with it!Slide34

Data available now

Data available very soon

Infected ash RNA-seq Illumina paired reads

Chalara genome sequence and gene annotationChalara

ITS sequenceChalara Calmodulin sequence

Ash genomic DNA

Illumina

paired reads

..your data?Slide35

Nornex – getting bigger

Lots of partners now agreeing to provide data and analyses on ash diebackSlide36
Slide37

What is the next step?

Continue to encourage engagement from experts

in the field to help with analyses

Oadb.tsl.ac.ukSlide38

MacLean Bioinformatics group

Dan MacLean

@danmacleanGraham Etherington

Kamoun Pathogenomics Group

Sophien Kamoun@kamounlabKentaro YoshidaDiane Saunders

Suomeng

Dong

Joe Win

University of Exeter

Genepool

(Edinburgh)

Forest Research

East Malling Research

Food and Environment Research Agency (FERA, York)

The John Innes Centre

The Genome Analysis Centre

University of Copenhagen

Norwegian Forest and Landscape InstituteSlide39

And you???

Oadb.tsl.ac.uk