+ - 0:00:00
Notes for current slide
Notes for next slide

Learning a Data-Driven Nosology:

Progress, Challenges & Opportunities

1 / 28

Steps



  1. Acquire data
  2. Build and run pipelines
  3. Develop and apply clustering methods
2 / 28

Progress: Datasets Acquired

3 / 28

Progress: Pipeline Built!

4 / 28

Clustering Methods: In Progress

5 / 28

Clustering Methods: In Progress

1) Don't find spurious clusters
2) Do find legitimate clusters

6 / 28

Challenge #1: Spurious Clusters

7 / 28

Challenge #1: Spurious Clusters

8 / 28

Challenge #1: Spurious Clusters

9 / 28

Challenge #1: Spurious Clusters

10 / 28

Challenge #1: Spurious Clusters

11 / 28

Challenge #1: Spurious Clusters

12 / 28

Challenge #1: Spurious Clusters

13 / 28

Challenge #1: Spurious Clusters

14 / 28

Opportunity #1: Spurious Clusters


Standard tools find spurious clusters
Fancy tools require too much data to be practical

Gap: Clustering tools with appropriate statistical guarantees

15 / 28

Challenge #2: Legitimate Clusters

16 / 28

Challenge #2: Legitimate Clusters

17 / 28

Opportunity #2: Legitimate Clusters


Thus far, harmonzing processing does not remove batch effect
Can further harmonizing processing mitigate batch effect?
If not, can harmonizing acquisition mitigate batch effect?

Gap: Pipelines that harmonize across datasets, and data to evaluate pipelines

18 / 28

Future Outlook

  • Microarrays never mitigated batch effects sufficiently
  • Neither microarrays nor MRI "count"
  • fMRI has another degree of freedom: stimulus
19 / 28

Future Outlook

  • Microarrays never mitigated batch effects sufficiently
  • Neither microarrays nor MRI "count"
  • fMRI has another degree of freedom: stimulus

Much More To Do!

(i love this guy)
20 / 28

Challenge #3: Overlapping Clusters

21 / 28

Challenge #3: Overlapping Clusters

22 / 28

Challenge #3: Overlapping Clusters

23 / 28

Challenge #3: Overlapping Clusters

24 / 28

Challenge #3: Overlapping Clusters

25 / 28

Challenge #3: Overlapping Clusters

26 / 28

Challenge #3: Overlapping Clusters

27 / 28

Opportunity #3: Overlapping Clusters


Any tree with thresholds will put arbitrarily close individuals arbitrarily far away


Trees that put individuals near boundary in multiple groups could mitigate this issue


Gap: Tree learning tools with appropriate statistical guarantees

28 / 28

Learning a Data-Driven Nosology:

Progress, Challenges & Opportunities

1 / 28

Steps



  1. Acquire data
  2. Build and run pipelines
  3. Develop and apply clustering methods
2 / 28

Progress: Datasets Acquired

3 / 28

Progress: Pipeline Built!

4 / 28

Clustering Methods: In Progress

5 / 28

Clustering Methods: In Progress

1) Don't find spurious clusters
2) Do find legitimate clusters

6 / 28

Challenge #1: Spurious Clusters

7 / 28

Challenge #1: Spurious Clusters

8 / 28

Challenge #1: Spurious Clusters

9 / 28

Challenge #1: Spurious Clusters

10 / 28

Challenge #1: Spurious Clusters

11 / 28

Challenge #1: Spurious Clusters

12 / 28

Challenge #1: Spurious Clusters

13 / 28

Challenge #1: Spurious Clusters

14 / 28

Opportunity #1: Spurious Clusters


Standard tools find spurious clusters
Fancy tools require too much data to be practical

Gap: Clustering tools with appropriate statistical guarantees

15 / 28

Challenge #2: Legitimate Clusters

16 / 28

Challenge #2: Legitimate Clusters

17 / 28

Opportunity #2: Legitimate Clusters


Thus far, harmonzing processing does not remove batch effect
Can further harmonizing processing mitigate batch effect?
If not, can harmonizing acquisition mitigate batch effect?

Gap: Pipelines that harmonize across datasets, and data to evaluate pipelines

18 / 28

Future Outlook

  • Microarrays never mitigated batch effects sufficiently
  • Neither microarrays nor MRI "count"
  • fMRI has another degree of freedom: stimulus
19 / 28

Future Outlook

  • Microarrays never mitigated batch effects sufficiently
  • Neither microarrays nor MRI "count"
  • fMRI has another degree of freedom: stimulus

Much More To Do!

(i love this guy)
20 / 28

Challenge #3: Overlapping Clusters

21 / 28

Challenge #3: Overlapping Clusters

22 / 28

Challenge #3: Overlapping Clusters

23 / 28

Challenge #3: Overlapping Clusters

24 / 28

Challenge #3: Overlapping Clusters

25 / 28

Challenge #3: Overlapping Clusters

26 / 28

Challenge #3: Overlapping Clusters

27 / 28

Opportunity #3: Overlapping Clusters


Any tree with thresholds will put arbitrarily close individuals arbitrarily far away


Trees that put individuals near boundary in multiple groups could mitigate this issue


Gap: Tree learning tools with appropriate statistical guarantees

28 / 28

Steps



  1. Acquire data
  2. Build and run pipelines
  3. Develop and apply clustering methods
2 / 28
Paused

Help

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
+ -
Notes for current slide
Notes for next slide
Paused

Help

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