SlideShare a Scribd company logo
Research is to see what everybody
else has seen and to think what
nobody else has thought.
- Albert Szent-Györgyi
Image by J.W. McGuire/NIH
Image from You Don’t Know Jack. Vol 3.
Data-driven pathway
analysis with ADAGE
Casey Greene
@GreeneScientist


Calvin and Hobbes. Bill Watterson
If you showed 16,000 computers 10 million
images from youtube, what would they see?
Le et al. 2012
Analysis with Denoising Autoencoders of
Gene Expression (ADAGE)
Tan et al. Pac Sym Bio 2015; Tan et al. mSystems 2016.
LeCun, Bengio, and Hinton. Nature 2015.
“Although we have not focused on it in
this Review, we expect unsupervised
learning to become far more important
in the longer term.”
Using ADAGE for pathway-style analyses
•  Pseudomonas aeruginosa compendium
•  > 100 different experiments
•  Many different labs
High-weight genes
HW relationships capture
genes’ pathways
Assign
Pathway
Tan et al. mSystems. 2016.
ADAGE analysis of publicly available
gene expression data collections
illuminates Pseudomonas aeruginosa-
host interactions
bioRxiv: http://dx.doi.org/10.1101/030650
github: http://github.com/greenelab/adage
Tan, Hammond, Hogan, and Greene. mSystems. 2016
Bigger models generally find
more pathways.
Tan et al. in prep.
The pathways captured by
ADAGE change with model size.
Tan et al. in prep.
We can measure gene-gene
similarity with ADAGE weights.
…
 …
ADAGE similarity captures
functional similarity.
Tan et al. in prep.
Image by Asafredell
ensemble ADAGE (eADAGE)
Tan et al. in prep.
eADAGE captures more pathways
than ADAGE.
Tan et al. in prep.
eADAGE captures more pathways
than ICA/PCA
Tan et al. in prep.
I didn’t want to just know the
names of things. I remember really
wanting to know how it all worked.
- Elizabeth Blackburn
Image: US Embassy Sweden
Activity volcano plot
Fold Change
-log(p)
Pareto activity selection
Fold Change
-log(p)
volcano + networks =
pathway-style analysis
ADAGE-based pathway analysis
reveals transcriptomic changes
Tan et al. in prep.
Where do we have (enough) data?
Greene et al. Pac Sym Bio. 2016
ADAGE webserver coming soon!
http://www.greenelab.com/webservers
Using ADAGE for pathway-style analyses
Research Parasite Awards
(The “Parasites”) 
Selection criteria for the work in question:
•  Not involved the design of the experiments that generated the data. 
•  Published independently of the original investigators
•  May have extended, replicated or disproved what the original investigators
had posited.
•  Provided source code and intermediate or final results in a manner that
enhances reproducibility.
Research Parasite Awards
(The “Parasites”) 
Additional selection criteria for the
Junior Parasite award:
•  The award is based on work described in a single manuscript.
•  Must have published the work at the training stage of their career
(postdoctoral, graduate, or undergraduate).
•  Should not have been in an independent investigator position for more than
2 years.
Research Parasite Awards
(The “Parasites”) 
Additional selection criteria for the
Sustained Parasitism award:
•  Must be an independent investigator in academia, industry or public sector.
•  Based on three manuscripts.
•  Must be last or corresponding author on each manuscript.
•  At least a five-year period must have elapsed between the publication of the
first manuscript and the final manuscript.
It ain’t what you don’t know
that gets you into trouble. It’s
what you know for sure that
just ain’t so.
- Attributed to Mark Twain
Greene Lab:
Jie Tan (Grad Student)
Gregory Way (Grad Student)
Brett Beaulieu-Jones (Grad Student)
René Zelaya (Programmer)
Matt Huyck (Programmer)
Dongbo Hu (Programmer)
Kathy Chen (Undergrad)
Mulin Xiong (Undergrad)
Tim Chang (Undergrad)

Collaborators:
Deb Hogan & Jack Hammond

Data:
All investigators who publicly release their gene
expression data.

Images:
Artists who release their work under a Creative
Commons license.

Funding:
Gordon and Betty Moore Foundation
National Science Foundation
Cystic Fibrosis Foundation
National Institutes of Health 
American Cancer Society

Find us online:
http://www.greenelab.com
Twitter: @GreeneScientist
Calvin and Hobbes. Bill Watterson

More Related Content

PDF
No Boundary Thinking in Bioinformatics Workshop Keynote
Casey Greene
 
PPT
Dr Julie Stahlhut - Barcode Data Life-cycle
Consortium for the Barcode of Life (CBOL)
 
PDF
Scouting_activity
kamwela oscar
 
PPTX
McIntosh "Improving the quality of preprints with automated checks"
National Information Standards Organization (NISO)
 
PPTX
SWAT4LS Open PHACTS Explorer demonstration
thetravellingbard
 
PPTX
MLA CE Course: Third-Party PubMed Tools
National Network of Libraries of Medicine, Pacific Northwest Region
 
PPTX
Pepe "Enriching Preprints with Provenance, Reproducibility, and Trustworthiness"
National Information Standards Organization (NISO)
 
No Boundary Thinking in Bioinformatics Workshop Keynote
Casey Greene
 
Dr Julie Stahlhut - Barcode Data Life-cycle
Consortium for the Barcode of Life (CBOL)
 
Scouting_activity
kamwela oscar
 
McIntosh "Improving the quality of preprints with automated checks"
National Information Standards Organization (NISO)
 
SWAT4LS Open PHACTS Explorer demonstration
thetravellingbard
 
Pepe "Enriching Preprints with Provenance, Reproducibility, and Trustworthiness"
National Information Standards Organization (NISO)
 

What's hot (20)

PDF
Why should Journals ask fo RRIDs?
Neuroscience Information Framework
 
PPTX
S17 biot6838 santiago
Katherine Magner
 
PDF
Casey Greene's Keynote for Rocky 2015
Casey Greene
 
PDF
Funk "Indexing & Discovering a Record of Versions"
National Information Standards Organization (NISO)
 
PPTX
Digital Scholarship and Open Science need a digital infrastructure
Björn Brembs
 
PPT
NEUR 1P01 winter 2019 ppt slides
Brock University
 
PPTX
Semantic approaches for biomedical knowledge discovery - Discovery Science 20...
Michel Dumontier
 
PPTX
How Can We Make Genomic Epidemiology a Widespread Reality? - William Hsiao
William Hsiao
 
PPTX
The infrastructure crisis of science
Björn Brembs
 
PDF
References on Reproducibility Crisis in Science by D.V.M. Bishop
Dorothy Bishop
 
DOCX
Bishop reproducibility references nov2016
Dorothy Bishop
 
PDF
2016 ACS Semantic Approaches for Biochemical Knowledge Discovery
Michel Dumontier
 
PPTX
IRIDA: Canada’s federated platform for genomic epidemiology
William Hsiao
 
PPTX
Biocuration 2014 - The Resource Identification Initiative
mhaendel
 
PPTX
Can machines understand the scientific literature
petermurrayrust
 
PPTX
Biovision2017 Accessing the scientific literature
petermurrayrust
 
PDF
Mozilla Science Labs Berlin Meetup
Brian Bot
 
PPT
Open Notebook Science in Drug Discovery
Jean-Claude Bradley
 
PPTX
Sc. fair research paper (2012 2013)
thompsonj1064
 
Why should Journals ask fo RRIDs?
Neuroscience Information Framework
 
S17 biot6838 santiago
Katherine Magner
 
Casey Greene's Keynote for Rocky 2015
Casey Greene
 
Funk "Indexing & Discovering a Record of Versions"
National Information Standards Organization (NISO)
 
Digital Scholarship and Open Science need a digital infrastructure
Björn Brembs
 
NEUR 1P01 winter 2019 ppt slides
Brock University
 
Semantic approaches for biomedical knowledge discovery - Discovery Science 20...
Michel Dumontier
 
How Can We Make Genomic Epidemiology a Widespread Reality? - William Hsiao
William Hsiao
 
The infrastructure crisis of science
Björn Brembs
 
References on Reproducibility Crisis in Science by D.V.M. Bishop
Dorothy Bishop
 
Bishop reproducibility references nov2016
Dorothy Bishop
 
2016 ACS Semantic Approaches for Biochemical Knowledge Discovery
Michel Dumontier
 
IRIDA: Canada’s federated platform for genomic epidemiology
William Hsiao
 
Biocuration 2014 - The Resource Identification Initiative
mhaendel
 
Can machines understand the scientific literature
petermurrayrust
 
Biovision2017 Accessing the scientific literature
petermurrayrust
 
Mozilla Science Labs Berlin Meetup
Brian Bot
 
Open Notebook Science in Drug Discovery
Jean-Claude Bradley
 
Sc. fair research paper (2012 2013)
thompsonj1064
 
Ad

Viewers also liked (15)

PPSX
Buhar ve Kızgın Yağ Kazanlarında Akışkan Yatak Teknolojileri
Deytamark Digital Pazarlama - SEO - Sosyal Medya
 
PPTX
Budgetseminar juni 2015
Claus Thykjær
 
PPTX
літопис 9 а
Vhitel
 
PPT
Narayaneeyam Tamil Dasakam 095
Ravi Ramakrishnan
 
PPTX
December 5 campus notes 12052013
Abigail Bacon
 
PDF
2016 Catalogue track and field 160318
George Chao (Sport)
 
PPTX
Tabajo de fotoshop
geiderlopez8-b
 
PPS
Presente De 15 Anos
Doni Cia
 
PDF
آموزش محاسبه انتگرال به کمک شبیه سازی مونت کارلو
faradars
 
PPTX
Apresentação bortoletto mz
Mazé Inácio
 
PPT
Ush ch. 22 section 2 notes
skorbar7
 
PDF
CV-ES Enslin
Eddie Enslin
 
PDF
Agility is not a dog show
Mark Nijssen
 
PDF
Seizures 60
gallevy16
 
Buhar ve Kızgın Yağ Kazanlarında Akışkan Yatak Teknolojileri
Deytamark Digital Pazarlama - SEO - Sosyal Medya
 
Budgetseminar juni 2015
Claus Thykjær
 
літопис 9 а
Vhitel
 
Narayaneeyam Tamil Dasakam 095
Ravi Ramakrishnan
 
December 5 campus notes 12052013
Abigail Bacon
 
2016 Catalogue track and field 160318
George Chao (Sport)
 
Tabajo de fotoshop
geiderlopez8-b
 
Presente De 15 Anos
Doni Cia
 
آموزش محاسبه انتگرال به کمک شبیه سازی مونت کارلو
faradars
 
Apresentação bortoletto mz
Mazé Inácio
 
Ush ch. 22 section 2 notes
skorbar7
 
CV-ES Enslin
Eddie Enslin
 
Agility is not a dog show
Mark Nijssen
 
Seizures 60
gallevy16
 
Ad

Similar to Using ADAGE for pathway-style analyses (6)

PDF
Public data as a lab resource.
Casey Greene
 
PDF
MBMWW2023_slideshare.pdf
Kazuya Horibe
 
PPT
Data at the NIH: Some Early Thoughts
Philip Bourne
 
PPTX
Scott Edmunds: Data Dissemination in the era of "Big-Data"
GigaScience, BGI Hong Kong
 
PPT
Some Early Thoughts
Philip Bourne
 
PDF
Revolutionizing medicine in the 21st century through systems approaches
Institute for Systems Biology
 
Public data as a lab resource.
Casey Greene
 
MBMWW2023_slideshare.pdf
Kazuya Horibe
 
Data at the NIH: Some Early Thoughts
Philip Bourne
 
Scott Edmunds: Data Dissemination in the era of "Big-Data"
GigaScience, BGI Hong Kong
 
Some Early Thoughts
Philip Bourne
 
Revolutionizing medicine in the 21st century through systems approaches
Institute for Systems Biology
 

Recently uploaded (20)

PPTX
Hydrocarbons Pollution. OIL pollutionpptx
AkCreation33
 
PDF
Migrating Katalon Studio Tests to Playwright with Model Driven Engineering
ESUG
 
PPTX
Brain_stem_Medulla oblongata_functions of pons_mid brain
muralinath2
 
PPTX
fghvqwhfugqaifbiqufbiquvbfuqvfuqyvfqvfouiqvfq
PERMISONJERWIN
 
PPTX
Internal Capsule_Divisions_fibres_lesions
muralinath2
 
PPTX
first COT (MATH).pptxCSAsCNKHPHCouAGSCAUO:GC/ZKVHxsacba
DitaSIdnay
 
PPTX
METABOLIC_SYNDROME Dr Shadab- kgmu lucknow pptx
ShadabAlam169087
 
PPTX
Reticular formation_nuclei_afferent_efferent
muralinath2
 
PPTX
Feeding stratagey for climate change dairy animals.
Dr.Zulfy haq
 
PDF
study of microbiologically influenced corrosion of 2205 duplex stainless stee...
ahmadfreak180
 
PPTX
Cell Structure and Organelles Slides PPT
JesusNeyra8
 
PDF
Multiwavelength Study of a Hyperluminous X-Ray Source near NGC6099: A Strong ...
Sérgio Sacani
 
PDF
JADESreveals a large population of low mass black holes at high redshift
Sérgio Sacani
 
PDF
A deep Search for Ethylene Glycol and Glycolonitrile in the V883 Ori Protopla...
Sérgio Sacani
 
PPT
1. Basic Principles of Medical Microbiology Part 1.ppt
separatedwalk
 
PPTX
Qualification of.UV visible spectrophotometer pptx
shrutipandit17
 
PPTX
Sleep_pysilogy_types_REM_NREM_duration_Sleep center
muralinath2
 
PPTX
Hepatopulmonary syndrome power point presentation
raknasivar1997
 
PPTX
Limbic system_components_connections_ functions.pptx
muralinath2
 
PDF
Approximating manifold orbits by means of Machine Learning Techniques
Esther Barrabés Vera
 
Hydrocarbons Pollution. OIL pollutionpptx
AkCreation33
 
Migrating Katalon Studio Tests to Playwright with Model Driven Engineering
ESUG
 
Brain_stem_Medulla oblongata_functions of pons_mid brain
muralinath2
 
fghvqwhfugqaifbiqufbiquvbfuqvfuqyvfqvfouiqvfq
PERMISONJERWIN
 
Internal Capsule_Divisions_fibres_lesions
muralinath2
 
first COT (MATH).pptxCSAsCNKHPHCouAGSCAUO:GC/ZKVHxsacba
DitaSIdnay
 
METABOLIC_SYNDROME Dr Shadab- kgmu lucknow pptx
ShadabAlam169087
 
Reticular formation_nuclei_afferent_efferent
muralinath2
 
Feeding stratagey for climate change dairy animals.
Dr.Zulfy haq
 
study of microbiologically influenced corrosion of 2205 duplex stainless stee...
ahmadfreak180
 
Cell Structure and Organelles Slides PPT
JesusNeyra8
 
Multiwavelength Study of a Hyperluminous X-Ray Source near NGC6099: A Strong ...
Sérgio Sacani
 
JADESreveals a large population of low mass black holes at high redshift
Sérgio Sacani
 
A deep Search for Ethylene Glycol and Glycolonitrile in the V883 Ori Protopla...
Sérgio Sacani
 
1. Basic Principles of Medical Microbiology Part 1.ppt
separatedwalk
 
Qualification of.UV visible spectrophotometer pptx
shrutipandit17
 
Sleep_pysilogy_types_REM_NREM_duration_Sleep center
muralinath2
 
Hepatopulmonary syndrome power point presentation
raknasivar1997
 
Limbic system_components_connections_ functions.pptx
muralinath2
 
Approximating manifold orbits by means of Machine Learning Techniques
Esther Barrabés Vera
 

Using ADAGE for pathway-style analyses

  • 1. Research is to see what everybody else has seen and to think what nobody else has thought. - Albert Szent-Györgyi Image by J.W. McGuire/NIH
  • 2. Image from You Don’t Know Jack. Vol 3.
  • 3. Data-driven pathway analysis with ADAGE Casey Greene @GreeneScientist Calvin and Hobbes. Bill Watterson
  • 4. If you showed 16,000 computers 10 million images from youtube, what would they see? Le et al. 2012
  • 5. Analysis with Denoising Autoencoders of Gene Expression (ADAGE) Tan et al. Pac Sym Bio 2015; Tan et al. mSystems 2016.
  • 6. LeCun, Bengio, and Hinton. Nature 2015. “Although we have not focused on it in this Review, we expect unsupervised learning to become far more important in the longer term.”
  • 8. •  Pseudomonas aeruginosa compendium •  > 100 different experiments •  Many different labs
  • 10. HW relationships capture genes’ pathways Assign Pathway Tan et al. mSystems. 2016.
  • 11. ADAGE analysis of publicly available gene expression data collections illuminates Pseudomonas aeruginosa- host interactions bioRxiv: http://dx.doi.org/10.1101/030650 github: http://github.com/greenelab/adage Tan, Hammond, Hogan, and Greene. mSystems. 2016
  • 12. Bigger models generally find more pathways. Tan et al. in prep.
  • 13. The pathways captured by ADAGE change with model size. Tan et al. in prep.
  • 14. We can measure gene-gene similarity with ADAGE weights. … …
  • 15. ADAGE similarity captures functional similarity. Tan et al. in prep.
  • 17. ensemble ADAGE (eADAGE) Tan et al. in prep.
  • 18. eADAGE captures more pathways than ADAGE. Tan et al. in prep.
  • 19. eADAGE captures more pathways than ICA/PCA Tan et al. in prep.
  • 20. I didn’t want to just know the names of things. I remember really wanting to know how it all worked. - Elizabeth Blackburn Image: US Embassy Sweden
  • 21. Activity volcano plot Fold Change -log(p)
  • 23. volcano + networks = pathway-style analysis
  • 24. ADAGE-based pathway analysis reveals transcriptomic changes Tan et al. in prep.
  • 25. Where do we have (enough) data? Greene et al. Pac Sym Bio. 2016
  • 26. ADAGE webserver coming soon! http://www.greenelab.com/webservers
  • 28. Research Parasite Awards (The “Parasites”) Selection criteria for the work in question: •  Not involved the design of the experiments that generated the data. •  Published independently of the original investigators •  May have extended, replicated or disproved what the original investigators had posited. •  Provided source code and intermediate or final results in a manner that enhances reproducibility.
  • 29. Research Parasite Awards (The “Parasites”) Additional selection criteria for the Junior Parasite award: •  The award is based on work described in a single manuscript. •  Must have published the work at the training stage of their career (postdoctoral, graduate, or undergraduate). •  Should not have been in an independent investigator position for more than 2 years.
  • 30. Research Parasite Awards (The “Parasites”) Additional selection criteria for the Sustained Parasitism award: •  Must be an independent investigator in academia, industry or public sector. •  Based on three manuscripts. •  Must be last or corresponding author on each manuscript. •  At least a five-year period must have elapsed between the publication of the first manuscript and the final manuscript.
  • 31. It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so. - Attributed to Mark Twain
  • 32. Greene Lab: Jie Tan (Grad Student) Gregory Way (Grad Student) Brett Beaulieu-Jones (Grad Student) René Zelaya (Programmer) Matt Huyck (Programmer) Dongbo Hu (Programmer) Kathy Chen (Undergrad) Mulin Xiong (Undergrad) Tim Chang (Undergrad) Collaborators: Deb Hogan & Jack Hammond Data: All investigators who publicly release their gene expression data. Images: Artists who release their work under a Creative Commons license. Funding: Gordon and Betty Moore Foundation National Science Foundation Cystic Fibrosis Foundation National Institutes of Health American Cancer Society Find us online: http://www.greenelab.com Twitter: @GreeneScientist Calvin and Hobbes. Bill Watterson