Identifying multimodal signatures underlying the somatic comorbidity of psychosis: the COMMITMENT roadmap

Emanuel Schwarz, Dag Alnæs, Ole A. Andreassen, Han Cao, Junfang Chen, Franziska Degenhardt, Daria Doncevic, Dominic Dwyer, Roland Eils, Jeanette Erdmann, Carl Herrmann, Martin Hofmann‐Apitius, Tobias Kaufmann, Nikolaos Koutsouleris, Alpha Tom Kodamullil, Adyasha Khuntia, Sören Mucha, Markus M. Nöthen, Riya Paul, Mads L. Pedersen, Andrés Quintero, Heribert Schunkert, Ashwini Kumar Sharma, Heike Tost, Lars T. Westlye, Youcheng Zhang, Andreas Meyer‐LindenbergMolecular 2020 · DOI · OpenAlex

Abstract

Individuals with schizophrenia face a life expectancy over a decade shorter than the general population, primarily due to somatic comorbidities like Type 2 diabetes (T2D) and cardiovascular disease (CVD). The COMMITMENT project (COMorbidity Modeling via Integrative Transfer machine learning in MENTal illness) addresses this "excess mortality" by proposing a computational roadmap to identify shared biological signatures. By leveraging multi-omic data—including genomics, epigenetics, and neuroimaging—and advanced transfer learning techniques, the project aims to move beyond traditional diagnostic silos. This paper outlines the framework for projecting disparate disease data into a shared biological subspace, enabling personalized risk stratification and preventative interventions.

Introduction: The Mortality Gap in Psychosis

A staggering 10-year mortality gap exists for individuals diagnosed with schizophrenia. This disparity is not solely a consequence of psychiatric symptoms but is driven by a high prevalence of physical health complications, notably T2D and CVD. Current clinical practice often treats these conditions in isolation, failing to account for the deep biological links—or pleiotropy—where shared genetic variants and pathways contribute to both mental and physical illness.

The COMMITMENT project seeks to bridge this gap by developing tools to stratify patients based on biological risk. By identifying which psychosis patients are predisposed to specific somatic complications, clinicians can shift from a "one-fits-all" approach to targeted, preventative care.

The COMMITMENT Framework: Multi-omic Integration

The core of the COMMITMENT roadmap is the integration of diverse "deep phenotyping" cohorts. These cohorts provide a wealth of multimodal data, including Genome-Wide Association Studies (GWAS), epigenetics, neuroimaging, clinical records, and gene expression data.

The COMMITMENT roadmap: Integrating deep phenotyping cohorts across psychosis and somatic case studies through multi-omic transfer learning to define a shared subgroup space.

As illustrated in Figure 1, the framework utilizes case studies across Psychosis, Type 2 diabetes, Cardiovascular disease, and Neurodegenerative conditions. By analyzing these cohorts in parallel, researchers can identify overlapping signatures that define a "Subgroup Space."

Methodological Innovations: Transfer Learning and NMF

To overcome the challenge of non-overlapping datasets—where not every patient has every type of data—the project employs two primary computational strategies:

  1. Transfer Learning: This approach allows researchers to identify signatures of a specific illness (e.g., T2D) in one specialized cohort and "transfer" those dimensions to psychosis cohorts. This enables the detection of somatic risk markers even in patients who have not yet developed physical symptoms.
  2. Non-negative Matrix Factorization (NMF): NMF is used to project multimodal data into a shared feature subspace. This mathematical framework is particularly effective at handling missing data points across different modalities, allowing for a unified analysis of genetics, imaging, and clinical records.

Furthermore, the framework integrates mechanistic biological knowledge to reduce data dimensionality and utilizes Transcriptome-wide Association Studies (TWAS) to predict how genetic variants influence tissue-specific gene expression, providing a causal link between genotype and phenotype.

Key Insights: Psychosis as a Systemic Disorder

The COMMITMENT project redefines schizophrenia as a systemic disorder rather than a purely localized brain condition. Key insights driving this roadmap include:

  • Independent Comorbidity: Metabolic abnormalities often appear in medication-naïve patients, suggesting that diabetes risk is an intrinsic part of the illness rather than just a side effect of antipsychotic medication.
  • Accelerated Brain Aging: Evidence of apparent brain aging in psychiatric patients suggests that neurodegenerative processes may serve as a link between mental health symptoms and physical decline.
  • Transdiagnostic Dimensions: By focusing on biological pathways rather than traditional diagnostic categories, the project aims to identify dimensions of illness that span multiple diagnoses.

Conclusion and Future Directions

The primary challenge for the COMMITMENT project lies in the complexity of fusing large-scale, distributed data. Future efforts will focus on mapping lifespan trajectories to identify critical age-periods for comorbidity risk and validating these machine-learning signatures in real-world clinical settings. Ultimately, this roadmap provides the foundation for a precision medicine approach to mental health, aiming to close the mortality gap through early, biologically-informed intervention.

Summary updated

Abstract

Developmental dyslexia (DD) is a learning disorder affecting the ability to read, with a heritability of 40-60%. A notable part of this heritability remains unexplained, and large genetic studies are warranted to identify new susceptibility genes and clarify the genetic bases of dyslexia. We carried out a genome-wide association study (GWAS) on 2274 dyslexia cases and 6272 controls, testing associations at the single variant, gene, and pathway level, and estimating heritability using single-nucleotide polymorphism (SNP) data. We also calculated polygenic scores (PGSs) based on large-scale GWAS data for different neuropsychiatric disorders and cortical brain measures, educational attainment, and fluid intelligence, testing them for association with dyslexia status in our sample. We observed statistically significant (p < 2.8 × 10?6) enrichment of associations at the gene level, for LOC388780 (20p13; uncharacterized gene), and for VEPH1 (3q25), a gene implicated in brain development. We estimated an SNP-based heritability of 20-25% for DD, and observed significant associations of dyslexia risk with PGSs for attention deficit hyperactivity disorder (at pT = 0.05 in the training GWAS: OR = 1.23[1.16; 1.30] per standard deviation increase; p = 8 × 10?13), bipolar disorder (1.53[1.44; 1.63]; p = 1 × 10?43), schizophrenia (1.36[1.28; 1.45]; p = 4 × 10?22), psychiatric cross-disorder susceptibility (1.23[1.16; 1.30]; p = 3 × 10?12), cortical thickness of the transverse temporal gyrus (0.90[0.86; 0.96]; p = 5 × 10?4), educational attainment (0.86[0.82; 0.91]; p = 2 × 10?7), and intelligence (0.72[0.68; 0.76]; p = 9 × 10?29). This study suggests an important contribution of common genetic variants to dyslexia risk, and novel genomic overlaps with psychiatric conditions like bipolar disorder, schizophrenia, and cross-disorder susceptibility. Moreover, it revealed the presence of shared genetic foundations with a neural correlate previously implicated in dyslexia by neuroimaging evidence.
BibTeX
@inproceedings{schwarz2020identifying,
  author = {Emanuel Schwarz and Dag Alnæs and Ole A. Andreassen and Han Cao and Junfang Chen and Franziska Degenhardt and Daria Doncevic and Dominic Dwyer and Roland Eils and Jeanette Erdmann and Carl Herrmann and Martin Hofmann‐Apitius and Tobias Kaufmann and Nikolaos Koutsouleris and Alpha Tom Kodamullil and Adyasha Khuntia and Sören Mucha and Markus M. Nöthen and Riya Paul and Mads L. Pedersen and Andrés Quintero and Heribert Schunkert and Ashwini Kumar Sharma and Heike Tost and Lars T. Westlye and Youcheng Zhang and Andreas Meyer‐Lindenberg},
  title = {{Identifying multimodal signatures underlying the somatic comorbidity of psychosis: the COMMITMENT roadmap}},
  booktitle = {Molecular},
  year = {2020},
  doi = {10.1038/s41380-020-00915-z},
}

Loading related content...

Figure