Using polygenic scores and clinical data for bipolar disorder patient stratification and lithium response prediction: machine learning approach.

2022-02-28

Cearns, Micah

Amare, Azmeraw T

Schubert, Klaus Oliver

Thalamuthu, Anbupalam

Frank, Joseph

Streit, Fabian

Adli, Mazda

Akula, Nirmala

Akiyama, Kazufumi

Ardau, Raffaella

Arias, Bárbara

Aubry, Jean-Michel

Backlund, Lena

Bhattacharjee, Abesh Kumar

Bellivier, Frank

Benabarre, Antonio

Bengesser, Susanne

Biernacka, Joanna M

Birner, Armin

Brichant-Petitjean, Clara

Cervantes, Pablo

Chen, Hsi-Chung

Chillotti, Caterina

Cichon, Sven

Cruceanu, Cristiana

Czerski, Piotr M

Dalkner, Nina

Dayer, Alexandre

Degenhardt, Franziska

Zompo, Maria Del

DePaulo, J Raymond

Étain, Bruno

Falkai, Peter

Forstner, Andreas J

Frisen, Louise

Frye, Mark A

Fullerton, Janice M

Gard, Sébastien

Garnham, Julie S

Goes, Fernando S

Grigoroiu-Serbanescu, Maria

Grof, Paul

Hashimoto, Ryota

Hauser, Joanna

Heilbronner, Urs

Herms, Stefan

Hoffmann, Per

Hofmann, Andrea

Hou, Liping

Hsu, Yi-Hsiang

Jamain, Stephane

Jiménez, Esther

Kahn, Jean-Pierre

Kassem, Layla

Kuo, Po-Hsiu

Kato, Tadafumi

Kelsoe, John

Kittel-Schneider, Sarah

Kliwicki, Sebastian

König, Barbara

Kusumi, Ichiro

Laje, Gonzalo

Landén, Mikael

Lavebratt, Catharina

Leboyer, Marion

Leckband, Susan G

Maj, Mario

Manchia, Mirko

Martinsson, Lina

McCarthy, Michael J

McElroy, Susan

Colom, Francesc

Mitjans, Marina

Mondimore, Francis M

Monteleone, Palmiero

Nievergelt, Caroline M

Nöthen, Markus M

Novák, Tomas

O'Donovan, Claire

Ozaki, Norio

Millischer, Vincent

Papiol, Sergi

Pfennig, Andrea

Pisanu, Claudia

Potash, James B

Reif, Andreas

Reininghaus, Eva

Rouleau, Guy A

Rybakowski, Janusz K

Schalling, Martin

Schofield, Peter R

Schweizer, Barbara W

Severino, Giovanni

Shekhtman, Tatyana

Shilling, Paul D

Shimoda, Katzutaka

Simhandl, Christian

Slaney, Claire M

Squassina, Alessio

Stamm, Thomas

Stopkova, Pavla

Tekola-Ayele, Fasil

Tortorella, Alfonso

Turecki, Gustavo

Veeh, Julia

Vieta, Eduard

Witt, Stephanie H

Roberts, Gloria

Zandi, Peter P

Alda, Martin

Bauer, Michael

McMahon, Francis J

Mitchell, Philip B

Schulze, Thomas G

Rietschel, Marcella

Clark, Scott R

Baune, Bernhard T

Mood stabilisers

bipolar affective disorders

depressive disorders

genetics

outcome studies

Journal Article

https://orcid.org/0000-0002-3353-8566

https://orcid.org/0000-0002-7940-0335

https://orcid.org/0000-0003-3191-0093

https://orcid.org/0000-0002-5377-1488

https://orcid.org/0000-0003-4987-2718

https://orcid.org/0000-0001-6219-4945

https://orcid.org/0000-0002-5287-3829

https://orcid.org/0000-0002-1966-5120

https://orcid.org/0000-0001-9544-3944

https://orcid.org/0000-0001-6548-426X

10.1192/bjp.2022.28

Response to lithium in patients with bipolar disorder is associated with clinical and transdiagnostic genetic factors. The predictive combination of these variables might help clinicians better predict which patients will respond to lithium treatment. To use a combination of transdiagnostic genetic and clinical factors to predict lithium response in patients with bipolar disorder. This study utilised genetic and clinical data (n = 1034) collected as part of the International Consortium on Lithium Genetics (ConLi+Gen) project. Polygenic risk scores (PRS) were computed for schizophrenia and major depressive disorder, and then combined with clinical variables using a cross-validated machine-learning regression approach. Unimodal, multimodal and genetically stratified models were trained and validated using ridge, elastic net and random forest regression on 692 patients with bipolar disorder from ten study sites using leave-site-out cross-validation. All models were then tested on an independent test set of 342 patients. The best performing models were then tested in a classification framework. The best performing linear model explained 5.1% (P = 0.0001) of variance in lithium response and was composed of clinical variables, PRS variables and interaction terms between them. The best performing non-linear model used only clinical variables and explained 8.1% (P = 0.0001) of variance in lithium response. A priori genomic stratification improved non-linear model performance to 13.7% (P = 0.0001) and improved the binary classification of lithium response. This model stratified patients based on their meta-polygenic loadings for major depressive disorder and schizophrenia and was then trained using clinical data. Using PRS to first stratify patients genetically and then train machine-learning models with clinical predictors led to large improvements in lithium response prediction. When used with other PRS and biological markers in the future this approach may help inform which patients are most likely to respond to lithium treatment.

link

The British Journal of Psychiatry : The Journal of Mental Science 2022; online first: 28 February

The British Journal of Psychiatry : The Journal of Mental Science