Jacob M Luber 1 2 Catherine Sharp 2 KB CHOI 2 Cheryl Ackertbicknell 2 3 amp MATThew a HIBBS 12 1 department of Computer Science Trinity University San Antonio ID: 534370
Download Presentation The PPT/PDF document "Predicting Functional Relationships in O..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Predicting Functional Relationships in Osteoblasts
Jacob M. Luber1, 2, Catherine Sharp2, KB CHOI2, Cheryl Ackert-bicknell2,3 & MATThew a. HIBBS1,21department of Computer Science, Trinity University, San Antonio, TEXAS 78212 USA2THE JACKSON LABORATORY, Bar Harbor, MAINE 04609 USA3University of Rochester medical center, Rochester, new york 14642 usaCorresponding author email: jluber@g.harvard.edu Slide2
Tissue Context Specificity
Bicknell & Hibbs, 2012Slide3
Functiona
l Relationship NetworksNode for each geneEdges between functionally related (or predicted genes)Correlation-based measures examine trends, rather than absolute valuesUnrelated pairs not connectedSlide4
Steps to Predict Improved Pathways
Mouse Biology Genomic DataFeaturesGold StandardHeterogeneous Data Integration Machine LearningPredictions!Slide5
Machine Learning & Context Specificity
We need to consider both: Tissue Specific Data (Bone Element)All Mouse Data
What Context Our
Data Come From
How We Handle
Ground Truth
&Slide6
ROC Curves
All Tissue ModelBone Only ModelGO Derived GSCurated GSMODEL TRAINED ON ALL TISSUE DATAWITH A MANUALLYCURATED GOLD STANDARDMODEL TRAINED ON ALL TISSUE DATAWITH A GO DERIVEDGOLD STANDARDMODEL TRAINED ON
BONE ELEMENT DATA
WITH A MANUALLY
CURATED GOLD STANDARD
MODEL TRAINED ON
BONE ELEMENT DATA
WITH A GO DERIVED
GOLD STANDARDSlide7
ROC Curves
All Tissue ModelBone Only ModelGO Derived GSCurated GSSlide8
PR Curves
All Tissue ModelBone Only ModelGO Derived GSCurated GSSlide9
WNT
Signaling (KEGG) Slide10
WNT Signaling
All Tissue ModelBone Only ModelGO Derived GSCurated GSSlide11
BMP Signaling (KEGG)
Slide12
BMP Signaling
All Tissue ModelBone Only ModelGO Derived GSCurated GSSlide13
Key Takeaways
Predictions made by the four classifiers are very dissimilarLikely that some of the highly predicted edges in classifiers trained on all data may not actually be related within the context of bone biology Literature evidence suggests classifiers trained on manually curated data and applied to only bone element data provides most accurate picture of bone biology (Cain et. al.)Curated gold standard contains edges not supported by bone only data---suggesting that only a subset of FRs in the literature are supported by co-expression data Methods are a next step of current state of the art methods like FNTMSlide14
Matt
HibbsCarol BultKB ChoiAdam Lavertu, Evan CoferCheryl Ackert-Bicknell, Catherine SharpTroyanskaya Group & Casey GreeneHuttenhower GroupNIH
NSF
Trinity
University Mach Research
Fellowship
More Details @
scholar.harvard.edu
/~
jluber
Contact Me @
jluber@g.harvard.edu
AcknowledgementsSlide15
Gaussian Fits