Nov.14.2017 9Pm CEST / 8Pm BST / 3Pm EDT / 12Pm PDT / 7Am GMT

1

ENIGMA-MDD

Nov.14.2017 | 9pm CEST / 8pm BST / 3pm EDT / 12pm PDT / 7am GMT

TELECONFERENCE |

─

Attendees

Lianne Schmaal

Paul Thompson

Agnes McMahon

Christopher Ching

Laura Nawijn

Laura van Velzen

Laura Han

Tracy Erwin-Grabner

Thomas Lancaster

Boris Gutman

Geoff Hall

Elena Pozzi

Evgeny Osipov

Dominik Grotegerd

Dario Zaremba

Philipp Saemann

Sinead Kelly

Meng Li

Tony Yang

Mat Harris

Nic van der Wee

Sean Hatton

Steven van der Werff

Lauren Salminen

Jennifer Evans

Andrian Campos

Thomas Frodl

Ian Harding

Lauren Salminen

Nynke Groenewold

Agenda

1. New members
2. Reminder to fill out data sharing approval for mega-analysis projects
3. Survey MDD Obesity and MDD Severity projects
4. New ENIGMA Suicidal Thoughts and Behaviours (STB) working group and NIH RO1 grant application
5. Updates projects
   

Action Items

1. Boris Gutman seeks help with interpreting and writing up the subcortical shape findings. Please contact Boris () with Lianne cc’d () if you are interested in this. This would be a great opportunity to be more involved without leading a project yourself!

1. Last chance to participate in the MDD Obesity, MDD Severity, MDD childhood trauma machine learning and MDD cortical suicidality projects! If you haven’t done so yet, please fill out the data sharing approval forms:

Obesity and Childhood trauma machine learning projects:

Severity project:

Cortical suicidality project:

1. If you are interested in participating in the new ENIGMA Suicidal Thoughts and Behaviors working group, please fill out the survey (if you haven’t done so yet):

1. Useful Links:
2. MOU:
3. Roadmap for initiating and leading an ENIGMA MDD meta- or mega-analysis projects:
   
4. Approved analysis plans and data sharing approval spreadsheets:
   
5. Google drive (pipelines, papers etc):
   
6. ENIGMA MDD website:
   

Business

‣New members:

○Jennifer Evans, Carlos Zarate from NIMH, NIH

○Kang Sim from Institute of Mental Health Singapore

○Welcome to these additional sites

‣For mega analysis projects:

○Some are underway

○Need sites approval in google form

○MDD Obesity and MDD Childhood Trauma Machine learning projects: (Missing: QTIM, Houston, Magdeburg, Barcelona)

○MDD Severity project: (Missing: MPIP, QTIM, Houston, Magdeburg)

‣Obesity and MDD severity projects

○Please fill out the surveys

○MDD Obesity project ():

(Missing: McMaster (Geoff Hall), Dublin (Thomas Frodl), Minnesota (Kathryn Cullen), MPIP (Philipp Saemann), Oxford (Beata Godlewska), EPISCA (Nic van der Wee))

○MDD Severity project ():
(Missing: several (16 of the 22 that filled out data sharing agreement form have filled out the survey))

‣Suicidal thoughts & behaviors

○Grant & new working group

■MQ Grant awarded to Lianne and colleagues: Charity foundation in UK

■NIH had an RFA for this as well - Neda & Lianne submitted grant

○Tony - adolescent suicidal thought and ideation - the MQ grant is specifically focused on this age group

■but the working group is all ages

■Transdiagnostic working group

■Participating sites in the new ENIGMA STB working group can also come up with secondary projects and come up with own analysis plans to lead projects

‣Project Updates

○MDD Lifespan (slide 13)

■Normative association between age and freesurfer brain measures, and how the MDD patients deviate

■Brain age gap - age as dependent variable (and brain measures as independent)

■Blue dots (hypothetical) would be healthy controls

■Then apply model parameters to patients to get predictive brain volume, for example (green dots)

■Get a z-score for each measure that represents difference between predicted and actual brain value (how much patients deviates from what would be expected on basis of their age)

■Associations for most brain measures are highly significant, done with gaussian process regression (captures also non-linear effects)

■In additional step we demonstrated that the associations were not confounded by Site/Scanner

■Looking at association with z-scores and clinical variables

●Some z-scores positive and some negative

●Maybe do a factor analysis, because so many z-scores (z-score for each subcortical and cortical freesurfer brain measure)

●Other option: do cluster analysis on Z scores and characterize resulting groups based on clinical measures

○Brain Age project

■Split into different training and test set, or by adolescent/adult → did not improve

■Controls into training set and test set, to see if predictive model generalized well

■Now using all controls and training set and MDD as test set

■Seems as though all machines (GPR, SVM, Random forest etc) perform quite similar

■Looking at difference in brain age gap between patients and controls and associations with clinical variables

■Will be writing up draft version to send around some time in december

■Slide 16 - best performance we’ve achieved so far

■Found some evidence for older brain age in adults with MDD, but gap is very small

○Severity questionnaires

○Cross-disorder project

■Bipolar-Scz-MDD

■Have agreed on additional variables that we will ask of the different samples

■Send around another email to confirm with SZ chairs that we can circulate those requests

○MDD DTI project

■Slide 21 (text below copied from slide)

■Total N= 1305 patients & 1602 controls

■Decrease FA (& increase in RD) in multiple regions, including the corpus callosum, corona radiata, internal capsule in adult MDD

■Decrease FA in the corpus callosum in adolescent MDD
( but does not survive FDR- correction)

■Strongest effects appear to be in adult MDD, recurrent MDD & in late age of onset

■Cohen’s D effect sizes similar to MDD cortical paper

■Performing some last analyses and hope to send the paper out soon

■Everyone involved thank you so much!

○Hippocampal Subfields project

■All samples together for final meta analysis

■Final count is 1911 controls, 3399 MDD patients

■Will try to send around results in next two weeks or so, and make suggestions on how to construct the paper

■Thanks to everyone who contributed so far

■(birds chirping in background)

○Shape project with Boris Gutman

■Meta analysis was largely finalized in march and April

■Diverse and interesting results, confirming volume subcortical findings, some extending previous subcortical findings

■Fixing some group issues

■You can have all these results, but it’s not as straightforward to write a story

■A couple things to make a good paper

●Boris could use some help from people who are more in the depression field, to help interpret the findings and write the paper

●Building quantitative interpretation mechanisms around these phenotypes

■Will start MDD paper with shape phenotypes soon, will need assistance with biological phenotype

■If you’re interested in helping Boris to write the paper: Reach out, cc’ing Lianne, with interest; great opportunity for active involvement

○Insomnia project

■Last conference call Eus presented results

■Will write amendment to alter proposed analyses and ask participating sites for permission for changes

○Machine learning

■Paper almost ready for submission to Molecular Psychiatry

■Brandy Riedel is taking over project from Daijang

■Classification performance was average

●Paper is about what factors contribute to variation in classification performance

●Interesting set of results

■Planning on submitting Thursday, if you received a link please fill out approval of manuscript, affiliations, acknowledgments etc

■Need approval from all co-authors before submitting, be very careful about that

■Paper will also be shared on biorxiv once submitted, can send around link to people who were not involved in this project who may be interested to read it

○Childhood adversity with Thomas Frodl

■Another paper that is almost ready

■Finishing discussion at the moment, will send around in week or two

■Very busy pre-christmas with all new manuscripts!

○Laterality project (slide 27)

■Clyde - update is that Tulio left his group and doesn’t have the time to work on this, another post-doc took over practicalities on project

■Waiting for DTA with SHIP, finalizing that agreement, then ready to do analyses

○MDD Obesity (slide 29)

■Update from Nils

■Lots of sites participating, those who filled out data sharing approval form

■Circulated survey to ask whether sites would want to share individual genetic data - most sites are unable/unwilling to do so

■Team is now starting with BMI data, will do that part of the analysis first

■Then see if there are alternative route to polygenic risk score analysis without sharing data

■Will be request shortly to ask for BMI measures

■Contact Nils directly with questions

○MDD Severity project

■Slide 32 - focus on some of results from survey

■Age of onset and number of episodes are already included in original MDD WG dataset

■Need to add additional covariates

■Slide 33 - preliminary study at Munster with voxel based data, and many clinical variables available

■Hospitalization factor scores from factor analysis, and found cortical and subcortical association

■Will send requests to participating sites for those additional covariates

■Can hopefully replicate analyses from Muenster in ENIGMA MDD

■Will start directly with data collection, fill out survey if you want but we are not waiting for it

○Childhood Trauma Machine Learning project

■Thanks to everyone’s participation on google form

■23 sites, will be reaching out to cohort PIs in next couple days

■Once coded we can start on pipeline and have some preliminary results for next call

■Now would be a good time to fill out google form if you’d like to participate!

■

○Suicide project (Miguel Renteria, via Lianne)

■22 groups filled out form

■Sent out survey to see who would want to participate in project, how suicidality was measured and whether sites would be interested in sharing data for a mega-analysis

■~2,905 MDD Cases with imaging; ~2,350 MDD cases with imaging + suicidality items; ~4,540 Healthy controls

■Severity of depression rating scale with questions about suicidal thoughts, but from more recent history, without lifetime attempts

■Assuming most sites will have clinical diagnostic interview; this would usually include info on the DSM MDD suicide criteria, including lifetime history of suicide attempts

■Ideally: Comparing 1) patients with suicidal ideation (no history of attempt), vs 2) patients with historyof attempt, vs 3) patients with no ideation and history of attempt, vs 4) healthy controls

■Should have large enough samples now (whereas when this was attempted a few years ago there weren’t yet) to make these comparisons

■Important and interesting to see what would differentiate attempters from ideators

■Slide 40:

●Proposed analysis plan, R scripts and timeline will be circulated among participating groups shortly.

●Possibility of running an updated meta-analysis of subcortical, ventricular and intracranial volumes.

■Mega analysis instead of meta analysis for cortical paper would require amendment to analysis plan and google form for data sharing approval

‣Conference calls

○2nd Tuesday of every month, every other month, at alternating times

○Will have to reschedule January’s for later in the month (was originally 9 January: 12 (noon) LA, 9 pm Amsterdam, 7 am (next day) Melbourne)

○13 March: 1 pm LA, 9 pm Amsterdam, 7 am (next day) Melbourne

○8 May: 6 am LA, 2 pm London, 3 pm Amsterdam, 11 pm Melbourne

○10 July: 6 am LA, 2 pm London, 3 pm Amsterdam, 11 pm Melbourne

‣Project participation tracking

○Elena trying to integrate information, so that there’s one central place to see where you’ve responded to what

○Will circulate when it’s ready, expect those emails over the next month

○Organic Data Science system will help with central organization eventually (through ENIGMA support group)

‣Thanks all for joining the call! Next call likely to be end of January