myGrid Workflow for Bioinformatics Analysis

http://taverna.org.uk/

S"an
Soiland-‐Reyes
&
Robert
Haines

myGrid,
School
of
Computer
Science

University
of
Manchester,
UK

ITER
IM
workshop

Château
de
Cadarache,
2011-‐06-‐08

What
is
myGrid?

  An
e-‐Science
Collabora"on
Since
2001

  Not
a
grid!

  Numerous
partners
involved:

  University
of
Manchester

  University
of
Southampton

  University
of
Oxford

  EMBL-‐EBI

  Provides
sustainable
and
produc"on
quality
soTware

  Supported
by
OMII-‐UK,
EPSRC
and
BBSRC

  Mixture
of
developers,
bioinforma"cians
and

researchers

SoTware
|
Services
|
Content
|
Skills
|
Community

http://www.mygrid.org.uk/
http://www.taverna.org.uk/

Mo"va"on:

Bioinforma)cs

  Challenge:

  Large
amounts
of
data

  Many
open
ques"ons

  Numerous
freely

available
public

datasets
and
analysis

tools


Huge
amounts
of
data

Microarray

1000+
Genes

QTL
regions

100+
Genes
How
do
I
look

Next
Gen

at
all
the
genes

systema)cally?
Sequencing

100,000+

Genes


Manual
approach

  Search
using
public
web
sites
and
databases

  Pubmed

  Uniprot

  EBI
BioMart

  Copy
and
paste
to
web
tools
for
analysis

  NCBI
Blast

  EBI
InterPro

  Further
processing
locally

  R

  Perl

  Python


Manual:
disadvantages

•  Scale
of
analysis
task
overwhelms
researchers

–
lots
of
data

•  User
bias
and
premature
ﬁltering
of
datasets
–

cherry
picking

•  Hypothesis-‐Driven
approach
to
data
analysis

•  Constant
changes
in
data
-‐
problems
with
re-‐
analysis
of
data

•  Implicit
methodologies
(hyper-‐linking
through

web
pages)

•  Error
prolifera)on
from
any
of
the
listed
issues

–
notably
human
error


Web
services
and
workﬂows

  Web
services

  Technology
and
standards
for
exposing
code
and

data
resources
that
can
be
programma)cally

consumed
by
a
remote
third
party

  Descrip"on
on
how
to
interact
with
the
service,

parameters,
documenta"on

  Workﬂows

  General
technique
for
describing
and
execu"ng

a
process

  Describe
what
you
want
to
do
running
which

services


The Taverna Open Source Suite of Tools Web Portals
Workflow Repository GUI Workbench Client User Interfaces

Virtual
Machine

Third Party Tools
Service Catalogue

Workflow Engine

Provenance Workflow
Store Server

Activity and Service
Plug-in Manager
Open
Provenance
Model

Programming and
Secure Service Access APIs

Taverna
workflows
Workflow Inputs

start_position chromosome_name end_position

genes_in_qtl

A
set
of
(local
and
remote)

mmusculus_gene_ensembl

remove_entrez_duplicates remove_uniprot_duplicates create_report
 
services
to
analyze
or
manage

merge_entrez_genes merge_uniprot_ids

remove_Nulls REMOVE_NULLS_2

data

add_ncbi_to_string add_uniprot_to_string

Kegg_gene_ids_2 Kegg_gene_ids

concat_kegg_genes

  Nested
workflows
are
also

split_gene_ids regex_2

split_for_duplicates

Get_pathways
remove_duplicate_kegg_genes

Workflow Inputs
services

Data-‐links
connects
services

regex gene_ids

split_by_regex

lister
 
get_pathways_by_genes1
  i.e.
output
from
service
A
is
input
to

service
B
and
C

Merge_pathways

concat_ids

  Describes
the
desired
dataflow

concat_gene_pathway_ids

Merge_gene_pathways

instead
of
process
coordina"on

Workflow Outputs

pathway_genes pathway_ids

merge_pathway_list_1

merge_pathway_list_2

split_for_duplicate_pathways
  Automa"c
itera"ons

Can
customize
list
handling
and

remove_duplicate_ids

pathway_descriptions
 
control
links

gene_descriptions merge_genes_and_pathways remove_pathway_duplicates

merge_gene_desc merge_genes_and_pathways_2 merge_pathway_desc

remove_nulls_3 merge_genes_and_pathways_3 remove_pathway_nulls merge_patwhay_ids species kegg_pathway_release

Workflow Outputs
flatten_pathway_files remove_pathway_nulls_2 merge_kegg_references merge_reports getcurrentdatabase binfo

gene_descriptions genes_pathways merged_pathways pathway_descriptions pathway_ids kegg_external_gene_reference report ensembl_database_release kegg_pathway_release


What
types
of
services
and
data?

  WSDL/SOAP
web
services

  Secured
invoca"on
with
HTTPS/SSL/WS-‐Security

  RESTful
web
services

  Secured
invoca"on
with
HTTPS/Basic
Auth

  Spreadsheet
import

  Command
line
tools
(local,
SSH)

  Inline
scripts
(Beanshell,
R)

  Excel/CSV
spreadsheets

  Java
APIs

  Customiza"ons:

  BioMart,
BioMoby
/
SADI

  Soaplab

  Grid
services
(EGEE
gLite,
caGrid,
PBS,
UNICORE)

  …
your
tool
(Plugin
tutorial
in
wiki)


Service
limita"ons

  Web
service
crea"on
involves
wrapping

exis"ng
tools
or
wri"ng
WS
code

  Web
services
can
go
down

  
can
use
redundant
services
in
workﬂow

  
Service
monitoring

  Transferring
data
up/down
to
WS
slow

  
Support
references
in
WS
interface

  Execu"ng
command
line
tools
directly
requires

execu"on
access

  Trickier
to
share
workﬂows,
require
either
SSH/grid

creden)als
or
installing
tools
locally


Which
services?

  Taverna
is
general,
can
connect
to
standard

web
services
and
command
line
tools
for
any

domain

  in
bioinforma"cs..

  From
professional
third-‐party
organisa"ons

providing
robust
&
open
data/analysis
services

  ..to
under-‐the-‐desk
web
services
for
one
par"cular

purpose,
ran
by
PhD
students

  
hhp://biocatalogue.org/
-‐
2000+
services
from

140+
providers
–
crowd
sourced
and
quality

monitored


BioCatalogue
integra"on

  Search
services
from

workbench

  Add
services
to
workﬂow

  View
service
descrip)ons

and
up)me
status
from

within
workﬂow


Taverna

workbench

  Graphical
desktop
tool

  No
server
installa"on

required

  Drag-‐and-‐drop
services

into
diagram

  Connect
services,
run,

reconnect,
rerun

  Integrates
diverse
set

of
tools


Sharing
workflows

  myExperiment.org
allows
users
to
share,

find,
download
and
rate
workflows

  “Facebook
for
the
scien"st”

  4000+
members,
1400+
workflows

  Open
source
code,
can
set
up
own
instance


myExperiment
integra"on

  Search
and
browse

workflows

  By
tags

  Free
text
search

  Own/group
workflows

  Packs,
e.g.
“Examples”

  Upload/share
workflows


Taverna
workﬂow
features

  Nested
workﬂows

  Reuse
exis"ng
components

  Implicit
itera"ons

  With
customizable
list
handling

  Pipelining

  Process
par"al
itera"on
results
early

  Parallelisa"on

  Run
as
soon
as
data
is
available

  Retries,
failover,
looping

  For
stability
and
condi"onal
tes"ng

  Plugin-‐extensible
execu"on
control

  Ideas:
caching,
error
detec"on,
dynamic
service
lookup


Extensible
UI
and
engine

  Plugins
can
provide
new
“perspec"ves”

  e.g.:
BioCatalogue,
myExperiment

  Provide
service-‐speciﬁc
customiza"on

  e.g.:
BioMart
interface
replicates
web
site

  Adding
new
func"onality

  New
service
types,
eg:
…

  Execu"on
control
like
looping/branching

  Design
helpers,
“Find
matching
service”


Workflow
limita"ons

  Ini"ally
designed
for
dataflows

  Not
suitable
for
business
processes
like
“HR

procedure
for
hiring
new
staff”

  Long-‐running
workflows
require
Taverna
Server

  ..
But
suitable
for
coordina)ng
command
line

and
grid
execu"ons,
the
data
might
just
be
job

references

  Execu"on
control
extensible,
eg:

  Looping,
Branching

  Dynamic
service
lookup

  Data
manipula"on,
Error
detec"on


Data
and
provenance
handling

  Data
references
passed
between
services
in
workflow

  http,
file,
sftp,
gridftp,
etc
(extensible)

  Data
downloaded/uploaded
or
references
translated

when
needed

  Provenance
captured
for
workflow
runs

  Trace
execu"on
steps,
view
intermediate
values
while
running

  Export
as
Open
Provenance
Model
(OPM)
/
RDF

  Proof
and
origin
of
produced
outputs

  Extensible
annota)ons

  Wf4Ever:
reproducible
research
objects

  Workflow/data
as
a
scien"fic
publica"on

preserva"on

  Need
to
capture
more
service
data
and
metadata


Data
limita"ons

  Running
Workbench
limited
by:

  Local
disk
space
for
storing
data

  Network
speeds
for
up/download

  Firewall
access

  
Execute
wf
using
Taverna
Server
or

command
line
remotely
with
ssh/job
submission

  No
standardized
WS
reference
mechanism

  Agree
on
mechanism
within
WS
‘family’
with

shared
disk
(eg.
deconstruct
local
path
from

HTTP
URI)


Parameter
sweeps

  Implicit
itera)ons
with
pipelining
provides

an
intui"ve
way
to
set
up
parameter

sweeps

  Advanced
looping
and
extensible
execu)on

control
allows
itera"ve
&
recursive

reduc"ons/approxima"ons


Taverna
command
line

  Executes
from
a

Windows/Linux/OSX

shells

  Takes
a
predefined

workflow
with
files
as

inputs
and
outputs

  Quick
way
to

“produc"onize”
a

workflow


Taverna
Server

  REST/SOAP
interface
to

execute
workﬂows

  Client
libraries
for
Ruby
and
Java

  Two
demonstra"on
web
interfaces

  Ruby

  Java
Portlets

  Upcoming:

  Security
delega"on

  AWS
image


Taverna
portlet

  Example
portlet

interface

  Executes
workﬂows

using
Taverna
Server


Ruby
web
interface

  Example
customized
  Uses
Ruby
gem

web
interface
t2-‐server


Grids
and
clusters

  Taverna
have
been
integrated
with
several

leading
grid
and
middleware

infrastructures,
such
as:

  PBS

  caGrid/Globus

  EGEE/gLite

  NorduGrid’s
ARC

  JSDL/GridSAM

  Plans
for
SAGA
integra"on


Taverna
on
the
cloud

  Use-‐case:

  SNP
analysis
and
annota"on
of

genome
sequenced
from

breeds
of
cows
in
Africa
–
why
are

some
of
them
resistent
to
X?

  Amazon
EC2
with
Taverna
Server
and
local

services

  Ruby
on
Rails
web
interface

  Runs
through
31
chromosomes
in
2
hours
using

10
instances
-‐
$10


Taverna
3
roadmap

  OSGi
plugin
system

  Workflow
language:
Scufl2

  Compound
format;
embedding
metadata,

dependencies,
independent
API
for
crea"ng/
inspec"ng
workflows

  Components

  Finding/sharing
command
line
tool
descrip"ons

  Richer
way
of
finding
compa"ble
services


Open
source,
open
development

  Taverna
suite
of
tools
are
all
open
source,

free
to
use
and
customize

  Large
user
community,
ac"ve
mailing
lists

  Lead
developers:
myGrid
in
Manchester
UK

  Contributors
from
across
the
world

  PAL
programme

  myGrid
provides
training,
tutorials
and

documenta)on


Who
uses
Taverna?

  Bioinforma"cs:
EMBL-‐EBI,
ONDEX

  Astronomy:
HELIO,
AstroGrid,
SAMPO

  Engineering:
NASA
Jet
Propulsion
Lab
(JPL)

  Chemistry:
CDK,
CIC

  Biodiversity:
BioVel

  Preserva"on:
Wf4Ever,
SCAPE

  BioMedicine/Cancer
research:
caGrid

  Data/text
mining:
eLico,
AID


Taverna
in
numbers

  myExperiment:

  4000+
registered
users

  56
countries

  Taverna:

  1400+
workﬂows

  361
organisa"ons

  48
countries

  BioCatalogue:

  70,000+
downloads

  2000+
services

  ~4000
source

  150+
service
providers

  500+
members

  27
countries


Acknowledgements


More
informa"on

  hhp://www.mygrid.org.uk/

  hhp://www.taverna.org.uk/

  hhp://www.myexperiment.org/

  hhp://www.biocatalogue.org/


myGrid Workflow for Bioinformatics Analysis

Recommandé

Recommandé

Contenu connexe

Similaire à myGrid Workflow for Bioinformatics Analysis

Similaire à myGrid Workflow for Bioinformatics Analysis (20)

Plus de Stian Soiland-Reyes

Plus de Stian Soiland-Reyes (17)

Dernier

Dernier (20)

myGrid Workflow for Bioinformatics Analysis