photo by Maurizio Pesce, CC-BY
Design of experiments
Design of experiments
Understanding the methods you use
Design of experiments
Understanding the methods you use
Case Studies
Design of experiments
Understanding the methods you use
Case Studies
Design of experiments
Understanding the methods you use
Case Studies
t-test
DESeq2
Simulations:
Simulations:
Easier
Test the whole process
Simulations:
Easier
Test the whole process
More assumptions
photo: U.S. government work
photo: U.S. government work
Cell culture
Does unoptanium increase midichlorian production?
Cell culture
Does unoptanium increase midichlorian production?
5 replicates
Cell culture
Does unoptanium increase midichlorian production?
5 replicates
Analyze with t-test, significant if p<0.05
Cell culture
Does unoptanium increase midichlorian production?
5 replicates
Analyze with t-test, significant if p<0.05
Simulation assumptions
+2μg
on average)Cell culture
Does unoptanium increase midichlorian production?
5 replicates
Analyze with t-test, significant if p<0.05
Simulation assumptions
Unoptanium helps ( +2μg
on average)
sd=8μg
Observed effect size
How frequently will we claim significance
Observed effect size
How frequently will we claim significance
Observed effect size
How frequently will we claim significance
a.k.a. power
But there's more!
Observed effect size
How frequently will we claim significance
a.k.a. power
But there's more!
Let's simulate 10000 datasets
photo: U.S. government work
Power:
## p < 0.05 in 0.0561 cases
Type S Error (wrong Sign)
Type S Error (wrong Sign)
Type S error | 95% CI excludes true |
---|---|
16.9% | 36.4% |
Type M Error (wrong Magnitude)
Type M Error (wrong Magnitude)
Mean exaggeration | Min. exaggeration |
---|---|
5.5 | 2.1 |
Published effects are exaggerated
Published effects are exaggerated
Exaggeration depends on amount of noise
Negligible in high-powered studies
Published effects are exaggerated
Exaggeration depends on amount of noise
Negligible in high-powered studies
If a results looks too good given the noise
Published effects are exaggerated
Exaggeration depends on amount of noise
Negligible in high-powered studies
If a results looks too good given the noise it probably is.
photo by Llann Wé, CC-BY
photo by Llann Wé, CC-BY
Differential expression upon unoptanium stress
Control, treatment, 3 replicates each
Differential expression upon unoptanium stress
Control, treatment, 3 replicates each
1000 genes
Differential expression upon unoptanium stress
Control, treatment, 3 replicates each
1000 genes
We use DESeq2 to test for effect = |log2(fc)|>1
Where do the read counts come from?
Where do the read counts come from?
How to set log2(fc)
?
Where do the read counts come from?
How to set log2(fc)
?
log2(fc)=0
Where do the read counts come from?
How to set log2(fc)
?
80% genes have log2(fc)=0
0, 2, 4 and 6 for the other 20%
Where do the read counts come from?
How to set log2(fc)
?
80% genes have log2(fc)=0
0, 2, 4 and 6 for the other 20%
100 simulations each
log_fc | True Pos. | False Pos. | Type S error | Mean exaggeration | Mean shrunk exaggeration |
---|---|---|---|---|---|
0 | 0.0 | 1.8 | 0.0 | NaN | NaN |
2 | 2.8 | 2.0 | 0.1 | 3.1 | 1.9 |
4 | 76.3 | 5.0 | 0.1 | 1.3 | 1.0 |
6 | 161.3 | 6.4 | 0.0 | 1.0 | 0.9 |
We tested for |log2(fc)|>1
Exact experiment replication (3 replicates each)
Replicated = significant in both
log_fc | Significant 1st experiment | Replicated | Smaller effect - significant |
---|---|---|---|
2 | 4.4 | 0.3 | 0.9 |
4 | 79.9 | 38.6 | 0.7 |
6 | 169.2 | 141.4 | 0.6 |
DE experiments have low power
DESeq2 rocks!
DE experiments have low power
DESeq2 rocks!
DESeq2 avoids false positives at all costs
DE experiments have low power
DESeq2 rocks!
DESeq2 avoids false positives at all costs -> high false negatives
Worry about Type S & M errors
Simulate experiments before investing money
Worry about Type S & M errors
Simulate experiments before investing money
Simulate to understand published research
Worry about Type S & M errors
Simulate experiments before investing money
Simulate to understand published research
Code available at https://github.com/cas-bioinf/statistical-simulations
Worry about Type S & M errors
Simulate experiments before investing money
Simulate to understand published research
Code available at https://github.com/cas-bioinf/statistical-simulations
Thanks for your attention!
log_fc | True Pos. | False Pos. | Type S error | Mean exaggeration | Mean shrunk exaggeration |
---|---|---|---|---|---|
0 | 0.0 | 0.7 | 0.0 | NaN | NaN |
2 | 8.1 | 0.8 | 0.0 | 1.8 | 1.4 |
4 | 150.9 | 2.4 | 0.1 | 1.1 | 1.0 |
6 | 184.1 | 3.4 | 0.0 | 1.0 | 0.9 |
We tested for |log2(fc)|>1
photo by Maurizio Pesce, CC-BY
Keyboard shortcuts
↑, ←, Pg Up, k | Go to previous slide |
↓, →, Pg Dn, Space, j | Go to next slide |
Home | Go to first slide |
End | Go to last slide |
Number + Return | Go to specific slide |
b / m / f | Toggle blackout / mirrored / fullscreen mode |
c | Clone slideshow |
p | Toggle presenter mode |
t | Restart the presentation timer |
?, h | Toggle this help |
Esc | Back to slideshow |