Transcript Real – time fMRI

Genotypes and phenotypes in Anorexia
Nervosa
A gene-association study
Outline
1. Genotypes and phenotypes in AN – general questions
2. Current works – analysis of 14 BMI-related SNPs
–
–
–
–
–
Genome-wide association study of BMI
Why to look at BMI-SNPs in ED?
Research questions
Results
Interpretation and conclusions
3. Bonus (if enough time) – article on real-time fMRI
Abbreviations
•
•
•
•
•
ED – Eating disorders
ANR – Anorexia nervosa restrictive type
GWAS – Genome wide association study
SNP – Single Nucleotide Polymorphism
BMI – Body mass index
General questions
• Are variations in particular genes associated with
susceptibility to ED?
• Are these genes associated with subsets of ED
patients?
• Are these genetic variations predictive of the
outcome and course of the treatment?
• To verify if variation in a gene is associated with a disease:
–
–
–
–
Genotype the candidate gene/polymorphism in cases and controls
Count genotypes in both groups
Compare frequencies of genotypes
By the way: Samples have to be large!
Current works – GWAS of BMI
• „Six New Loci Associated with BMI Highlight a
Neuronal Influence on Body Weight Regulation” –
Willer et al., Nature Genetics, January 2009
• Metaanalysis of previous findings plus replication with
new samples (~88k subjects)
• Unfortunately, it wasn’t done by us 
• Confirmed previous loci (FTO, MC4) and identified
new ones (p<5 x 10-8)
• 14 SNPs all together associated with BMI
14 SNPs associated with BMI in normal
population
•
•
•
•
•
•
•
•
•
•
•
FTO
MC4R (2 SNPs)
BDNF (2 SNPs)
TMEM18
GNPDA2
SH2B1
KCTD15
MTCH2
NEGR1
ETV5
2 SNPs not close to any gene
The Idea!
• Take these 14 SNPs associated with BMI in normal
population
• Genotype them in the sample of ED and controls
• Analyze – show association or lack of it
• With the results, address some of the questions
perturbing ED field
• ED types can be divided in a couple of subcategories
• Here, we focus only on one subcategory: ANR
Why to look at these SNPs - Questions
• In the ED field:
– Is high BMI a protective factor against ANR?
– Is ANR on the extreme end of the BMI dimension or is it
qualitatively different?
• In the current study:
– Are variants associated with high BMI in normal population
underrepresented in ANR?
– Are these variants associated with BMI in ANR sample
(Is low BMI in ANR etiologically distinct from low BMI in
normal population?)
Materials and methods
• A sample
– 205 female ANR patients and 1674 female healthy controls
(population sample)
– Some removed for low genotyping (more than 1 missing genotype)
– Resulting sample consisted of 173 ANR and 1571 control
subjects
• SNPs
– rs2844479 excluded because of significant differences in
missingess rate bt. cases and controls
– Remaining 13 SNPs passed the quality check
• Analyses were performed with Plink, UNPHASED, SPSS
Results: case-control
• To find out if SNPs are associated with ANR caseness
• A case-control analysis of frequencies of genotypes and
alleles showed no difference bt. groups.
Results: counting effect alleles
• To find out if BMI-increasing alleles are protective against ANR
• How to do it?
1. Make sure which allele in the GWAS study was the BMI-increasing
allele (effect allele) in each SNP
2. Calculate how many of effect alleles each individual in the current
study has
3. Compare means bt. cases and controls
Counts
Mean nr of
effect alleles
F
Sig.
Controls
1208
12.437
0.002
0.963
ANR Cases
(only complete genotypes)
131
12.427
• No difference in number of effect alleles bt. cases and controls
Results: BMI in ANR
• To find out if investigated SNPs affect BMI in ANR
sample
• rs925946 (BDNF gene) significantly associated with
BMI in ANR sample, p-value=0.008
Genotypes
Count
Freq.
Additive Value
Conf. Int.
95%Low
G/G
79
48.47%
0
0
0
G/T
68
41.72%
-0.004
-0.048
0.040
T/T
16
9.82%
0.188
0.012
0.364
• No association for other SNPs
Conf. Int.
95%High
Summary and discussion
• The aim of this study was to find out whether genetic variants
affecting BMI in normal population are protective against ANR
• Results show that this is not the case
• Additionally, except for one, SNPs influencing BMI in normal
population didn’t predict BMI in the ANR group, suggesting that
their effects are overruled by presence of other, putative
genetic risk factors for ANR
• Genetic etiology of low BMI in ANR is distinct from that in
normal population
• It suggests that ANR is qualitatively different from other types
of weight related disorders
Teaser
• There are also interesting results for the other
subcategory of ED
• 4 SNPs associated with bingeing-purging category
(ANP + BNP + NAO-P)
• That’s another story
Acknowledgements
• Roger Adan
• Judith Hendriks
• Annemarie van Elburg
• Unna Danner
The Research Training Network INTACT
Real – time fMRI
Traditional fMRI
– BOLD signal (Blood-oxygen-level dependent)
– Brain activity via oxygen absorption from hemoglobin (change
in magnetic properties)
– Delay: detectable oxygenation change after 3-5 seconds
– Takes time to analyse
Real-time fMRI
• Sophisticated software to analyse BOLD signal onthe-fly
• Real-time chagnes of activity observable (still with 35s delay)
• Visualized on the screen
• Adaptive experiments
Neurofeedback
• Improved biofeedback…
• Voluntary modulation of brain activity in chosen
regions
• Quick to grasp
• Subjects choos their own strategy and region (what
works best)
Some applications
• Brain-pong (controlling external devices)
• Locked-in syndrome (brain injury, stroke, ALS)
– Efficient communication, Brain-read
•
•
•
•
Pain perception
Empathy in children (learning brain states)
Lie detection (possibly)
Therapeutic applications?
– Stroke rehabilitation, epilepsy, depression, addiction treatment
– Augmentation of psychotherapy
• Possible implications… let loose your imagination
Caveats
• Many limitations
– Practical
– Neuroethics
– Still in its infancy
• Nature Reviews Neuroscience – „Applications of realtime fMRI”, deCharms, 2008