Talk presented at the WCCI in Vancouver/Canada, 2016

1. Gargi Mukherjee … Rutgers University, New Jersey
Kevin Raines … Stanford University, California
Srikanth Sastry … JNC, Bengaluru, India
Sebastian Doniach … Stanford University, California
Gyan Bhanot … Rutgers University, New Jersey
Michael Biehl … University of Groningen, The Netherlands
1
Predicting Recurrence in Clear Cell
Renal Cell Carcinoma
Analysis of TCGA data using Outlier Analysis and GMLVQ

2. WCCI 2016, Vancouver / BC 2 /15
overview
gene expression in tumor cells
specific example: clear cell Renal Cell Carcinomas (ccRCC)
• outlier analysis: identification of a panel of prognostic genes
with respect to recurrence
• risk score: prediction of individual recurrence risk
based on outlier status w.r.t. selected genes
• machine learning: analysis of extreme cases of low / high risk
distance based classification and relevance learning
(Generalized Matrix Relevance LVQ)
clinical data: recurrence free intervals

3. WCCI 2016, Vancouver / BC 3 /15
clear cell Renal Cell Carcinoma (ccRCC)
publicly available datasets:
The Cancer Genome Atlas (TCGA) cancergenome.nih.gov
also hosted at Broad Institute gdac.broadinstitute.org
data

4. WCCI 2016, Vancouver / BC 4 /15
data
20532genes
65normalsamples
469 tumor
samples
65 + 65
matched
clear cell renal cell carcinoma
TCGA data @ Broad Institute
mRNA-Seq expression data X
normalized, log-transformed:
Y=log(1+X)
65 normal samples
65 matched tumor samples
469 tumor samples in total
number of
recurrences
recurrence data:
days after diagnosis

5. WCCI 2016, Vancouver / BC 5 /15
380
training
samples
outlier analysis
89testsamples
randomized split

6. WCCI 2016, Vancouver / BC 6 /15
380
training
samples
outlier analysis
per gene:
determine
mean μ, standard deviation σ of Y
for each gene: identify outlier samples
Y > μ + σ “high outlier“
Y < μ - σ “low outlier“
restrict the following analysis to genes with
≥ 20 high outlier samples
or ≥ 20 low outlier samples

7. WCCI 2016, Vancouver / BC 7 /15
1546 „high-outlier genes“
with KM log rank p < 0.001
1628 „low-outlier genes“
with KM log rank p < 0.0005
construct two binary outlier matrices
„1“ for high-outlier samples
„0“ else
„1“ for low-outlier samples
„0“ else
1546 genes
 PCA
Kaplan-Meier (KM) analysis per gene:
test for significant association of outlier status of samples with recurrence
outlier analysis
1628 genes
380samples380samples

8. WCCI 2016, Vancouver / BC 8 /15
PCA reveals
four clusters of genes
711475
2261402
A B
DC
high outlier genes
low outlier genes
genes in small clusters (B,D):
outlier status associated
with late recurrence
genes in large clusters (A,C):
outlier status associated
with early recurrence
outlier analysis

9. WCCI 2016, Vancouver / BC 9 /15
recurrence risk score
top 20 genes (by KM p-value) from each cluster A,B,C,D
reference set of 80 genes
for each sample:
- determine outlier status with respect to the 80 genes (Y >?< μ ± σ )
- add up contributions per gene
- 1 if the sample is outlier w.r.t. to a gene in A or C (early rec.)
0 if the sample is not an outlier w.r.t. the gene
+ 1 if the sample is outlier w.r.t. to a gene in B or D (late rec.)
recurrence risk score - 40 ≤ R ≤ + 40
observe: median = 2 over the 380 training samples
crisp classification w.r.t. recurrence risk:
high risk (early recurrence) if R < 2
low risk (late recurrence) if R ≥ 2

10. WCCI 2016, Vancouver / BC 10 /15
recurrence risk prediction
training set (380 samples) test set (89 samples)
log rank p < 1.e-16 log rank p < 1.e-4
KM plots with respect to high / low risk groups:
• risk score R is predictive of the actual recurrence risk
• the 80 selected genes can serve as a prognostic panel

11. WCCI 2016, Vancouver / BC 11 /15
extreme case analysis
number of
recurrences:
≤ 2 years
(early)
> 5 years
(late or no
recurrence)
109 samples
class 2, high risk
107 samples
class 1, low risk
(undefined)
2 classes:
• 80-dim. feature vectors (gene expression)
• representation by one prototype vector per class:
• adaptive distance measure for comparison of samples and prototypes:
with relevance matrix
• distance-based classification, e.g. Nearest Prototype Classifier (NPC)

12. WCCI 2016, Vancouver / BC 12 /15
GMLVQ classifier
Generalized Matrix Relevance Learning Vector Quantization (GMLVQ)
training of prototypes and relevance matrix
= minimization of an appropriate cost function
with respect to performance on labeled training set
components of diagonal elements of Λ
A B C D A B C D
lowexpression|highexpression

13. WCCI 2016, Vancouver / BC 13 /15
GMLVQ classifier
ROC of GMLVQ classifier (Leave-One-Out of the 216 extreme samples)
KM plot w.r.t. all 469 samples
( L-1-O for 216 samples, plus 253 undefined )
log rank p < 1.e-7

14. WCCI 2016, Vancouver / BC 14 /15
extreme case analysis (107+109 samples)
GMLVQ classifier Risk score classifier
- AUC=0.84
 R=2

15. WCCI 2016, Vancouver / BC 15 /15
the set of 80 genes is also diagnostic:
• GMLVQ separates normal from tumor cells (close to) perfectly
• PCA of corresponding gene expressions:
65 normal samples
105 low risk samples (late recurrence)
109 high risk samples (early recurrence)
gradient from normal to high risk:
diagnostics?

16. WCCI 2016, Vancouver / BC 16 /15
• GMLVQ suggests an even smaller panel of prognostic genes (12?)
identify a minimum panel for diagnostics and prognostics
• 80 genes do not necessarily reflect biological mechanisms
compare, e.g., with known pathways / modules of genes
remarks and open questions
• prospective studies required with respect to use as an assay
• can the performance be improved further ?
study more sophisticated classifier systems
include further clinical information (diet, life style, family history, … )
easy-to-use GMLVQ-classifier: www.cs.rug.nl/~biehl/gmlvq
• more direct, multivariate identification of relevant genes ?
e.g. PCA+GMLVQ and back-transform