Transcript A new way of ordering endophenotypes for relevance to a disease

The search process:
A way to rank endophenotypes
(e) for relevance to a disease (i)
phenotype
“What is an endophenotype?” you
ask.
Great Question.
Some Candidates.
THESE WERE PLUCKED OUT OF 11,000 TRAITS:
Behavioral/Neurocognitive
Beck Depression Inventory II
EPQ Neuroticism
Declarative Memory (CVLT
Recognition) Working Memory (Digit
Span Forward)
Working Memory (Letter-Number
Span)
Penn Facial Memory (Immediate)
Penn Facial Memory
(Delayed)Attention (CPT hits) Attention
(Trails A) Penn Emotion Recognition
Neuroimaging
Ventral diencephalon volume Parietal
hyperintensity volume Hippocampus
volume Pallidum volume Cerebellar
white matter volume Frontal
hyperintensity volume Corticospinal
tract (FA) Subcortical hyperintensity
volume Superior parietal gyrus
thickness Thalamus proper volume
Transcriptional
RNF123 (3p24) PDXK (21q22) ZFP64 (20q13) RWDD2A
(6q14) B4GALT7 (5q35) MARK2 (11q12) GADD45A (1p31)
PIGN (18q21) HTT (4p16) ABHD12 (20p11)
Glahn et al., 2012 High Dimensional Endophenotype Ranking in the Search for Major Depression Risk Genes
Endophenotypes are:
•
•
•
•
NOT A CLINICAL SUBTYPE
HERITABLE
Trait-related deficit
An INTERMEDIATE phenotype that stresses the
genetic link
• State-INDEPENDENT (present whether or not
disease is active)
• Found WITH GREATER FREQUENCY (as compared
to general population) in related members of
affected individuals and affected individuals.
• The idea is that these will be less heterogeneous
than clinical measures
Braff et al., 2007 Deconstructing Schizophrenia
Gottesman, 2003 Psychiatric Endophenotypes and the development of valid animal models
Endophenotype Ranking Value (ERV):
An Index of genetic utility (standardized genetic
covariance) of e for i
ERVie 
2
i
h
h g
2
e
i = illness
e= endophenotype
ρg= genetic correlation of i & e
About the Endo Ranking Value
• ERV’s will vary between 0 and 1
– higher values ≈ stronger shared genetic influence
• ERV balances the strength of the genetic
signal for the e and the strength of its relation
to the i
• Advantage: large #s of potential es can be
assessed before conducting molecular genetic
analyses- useful in guiding the search.
ERV ingredients
ERVie 
h 
2
 p  g
 g2

2
p
 additive
g 
 g(i, e)
e
2
e
 g (i) g (e)
h
h g
2
e
Broad genetic heritability is the genotypic
variance divided by the phenotypic variance
h h   1 h 1 h 
2 2
e i
2
i
2
i
Phenotypic correlation derived from the
genetic and environmental correlations
(obtained by using all pedigree info avail with
a multivariate normal threshold model for
combined dichotomous/continuous traits
Let’s take a look at what an ERV looks like when you change the heritability values of e and i…
The Example Study: recurrent Major
Depressive Disorder (rMDD)
• N= 1122 from 71 extended pedigrees
• 11,000 traits
• FIRST: they rank their endophenotypes
• SECOND: they check the utility of these
rankings with bivariate linkage analyses (rMDD
+ endophenotypes)
Bivariate Linkage Analysis (to improve
power to detect)
• Exploits the genetic covariance between the endophenotype and
the illness to improve the power to localize QTLs (quantitative trait
locus) and to detect QTL-specific pleiotropic effects
• Compute maximum likelihood estimates of allele frequencies
• Matrices of empirical estimates of identity-by-descent allele sharing
at points throughout the genome for every relative pair were
computed
• Once genome-wide significant localization was made, formal single
degree of freedom likelihood ratio tests for pleiotropy were
performed to test the specific hypothesis that a QTL at that location
influenced a given endophenotype risk.
QTLs (quantitative trait loci) found
• 4p15 (using univariate linkage)
– Bilateral ventral p= 4.7 x 10-5
– RNF123 p= 0.0010
– Diencephalon volume p= 0.0290
– BDI-II p= 0.1170 (trend)
– The linkage analysis provides further validation
that the endophenotype identification strategy
can be used to find genetic influence
– Now these can be molecularly analysed.
Testing for Pleiotropic Effects
THESE WERE PLUCKED OUT OF 11,000 TRAITS:
Top Results
Behavioral/Neurocognitive
Beck Depression Inventory II
EPQ Neuroticism
Declarative Memory (CVLT
Recognition) Working Memory (Digit
Span Forward)
Working Memory (Letter-Number
Span)
Penn Facial Memory (Immediate)
Penn Facial Memory
(Delayed)Attention (CPT hits)
Attention (Trails A) Penn Emotion
Recognition
Neuroimaging
Ventral diencephalon volume Parietal
hyperintensity volume Hippocampus
volume Pallidum volume Cerebellar
white matter volume Frontal
hyperintensity volume Corticospinal
tract (FA) Subcortical hyperintensity
volume Superior parietal gyrus
thickness Thalamus proper volume
Transcriptional
RNF123
(3p24)
PDXK (21q22)
ZFP64 (20q13) RWDD2A (6q14) B4GALT7 (5q35) MARK2
(11q12) GADD45A (1p31) PIGN (18q21) HTT (4p16)
ABHD12 (20p11)
Glahn et al., 2012 High Dimensional Endophenotype Ranking in the Search for Major Depression Risk Genes
Varying heritability affecting the ERV
value.
ht Mark + Tim