The Madrid Manic Group (MadManic) Cohort: Multi-Omics and Digital Phenotyping For the Studies of Severe Mental Disorders and Suicidality

The Madrid Manic Group (MadManic) Cohort is a large-scale Spanish initiative integrating multi-omics, clinical, and digital phenotyping data from over 4,400 participants to investigate the biological basis of severe mental disorders and suicidality, thereby advancing precision psychiatry and enhancing the representation of Southern European populations in global psychiatric research.

The Gist

Imagine trying to solve a massive, tangled knot of string that represents severe mental illnesses like bipolar disorder, schizophrenia, and severe depression. For decades, scientists have been pulling on single strands of the knot, trying to figure out why it's so tight and why it sometimes leads to tragic outcomes like suicide.

The paper you're reading introduces a new, massive toolkit called the MadManic Cohort (or "MadManic" for short). Think of this not just as a study, but as a giant, high-tech detective agency based in Madrid, Spain, dedicated to untangling that knot.

Here is how they are doing it, broken down into simple concepts:

1. The "Super-Database" (The Cohort)

Imagine a library that doesn't just hold books, but holds the biological blueprints of over 4,400 people.

• The Cast: About 2,300 of these people are patients struggling with severe mental health issues, and 2,100 are healthy volunteers (the "control group").
• The Treasure: For each person, the team has collected a "biological time capsule." This includes blood samples (which give them DNA, RNA, and plasma) and saliva samples. It's like having a full instruction manual for how their bodies and brains are built.

2. The "Digital Shadow" (Digital Phenotyping)

Usually, doctors ask patients, "How are you feeling?" once a month. But feelings change by the hour.

• The Analogy: Imagine if, instead of asking a patient once a month, you gave them a smartwatch that quietly watches their life 24/7.
• How it works: The MadManic team uses two smartphone apps. One asks active questions (like "How is your sleep?"), and the other acts as a passive sensor, tracking how much they move, how fast they type, their voice tone, and even how far they walk.
• The Goal: This creates a "digital shadow" of the patient's life. It helps doctors see patterns they would never catch in a standard office visit, like noticing that a patient's sleep pattern changes two days before a manic episode hits.

3. The "Three-Layer Cake" (Multi-Omics)

To understand the disease, the team isn't just looking at one thing; they are looking at three different layers of biology at the same time:

• Layer 1: The Blueprint (Genomics): They look at the DNA to find the "hardwired" instructions. It's like checking the factory manual to see if there are typos in the original design.
• Layer 2: The Factory Floor (Transcriptomics): They look at RNA. If DNA is the blueprint, RNA is the construction crew actually building things. This tells them which parts of the blueprint are being used right now.
• Layer 3: The Switches (Epigenomics): They look at methylation. Think of this as sticky notes on the blueprint that say "Turn this off" or "Turn this on." These notes change based on life experiences, stress, and environment.

4. The "Why" and the "How"

Why do all this?

• The Puzzle: Mental illness is like a complex puzzle where the pieces are genetics, environment, and daily habits. By stacking all these layers (DNA + RNA + Digital Data), the team hopes to see the whole picture.
• The Goal: They want to move away from "one-size-fits-all" medicine. Instead of guessing which pill might work, they want to predict exactly which treatment will work for your specific biological makeup.
• The Suicide Mission: A major focus is preventing suicide. By combining the "digital shadow" (seeing a patient is acting differently) with the "biological blueprint" (seeing they have high genetic risk), they hope to build an early warning system to save lives.

5. The "Missing Piece" (Southern Europe)

Most of the world's big genetic studies have been done on people of Northern European descent. It's like trying to understand how all cars work by only studying Fords.

• The Contribution: The MadManic cohort is filling a huge gap by focusing on Southern Europeans (Spanish people). This ensures that the medical breakthroughs they find will actually work for a wider range of people, not just a specific group.

In a Nutshell

The MadManic Cohort is a massive, high-tech effort to stop guessing when it comes to mental health. By combining DNA, real-time smartphone tracking, and advanced lab tests, they are building a super-powered map. This map will help doctors navigate the complex terrain of mental illness, predict crises before they happen, and prescribe the right medicine for the right person, turning the tide against severe mental disorders and suicide.

Technical Summary

1. Problem Statement

Severe Mental Disorders (SMDs), including Bipolar Disorder (BD), Schizophrenia (SCZ), and Major Depressive Disorder (MDD), represent a significant global health burden characterized by high rates of suicide, functional impairment, and treatment resistance. Current clinical approaches face several limitations:

• Diagnostic Heterogeneity: Traditional diagnostic instruments rely on cross-sectional, face-to-face assessments that lack the granularity to predict individual risk trajectories or treatment responses.
• Genetic Complexity: While Genome-Wide Association Studies (GWAS) have identified numerous risk loci, these findings often lack functional context. Furthermore, large-scale genomic studies have historically underrepresented Southern European populations, limiting the generalizability of Polygenic Risk Scores (PRS) and precision medicine applications.
• Lack of Integration: There is a disconnect between genomic data, multi-omics layers (transcriptomics, epigenetics), and real-time clinical phenotyping. Existing cohorts often fail to integrate digital monitoring with deep molecular profiling, hindering the development of predictive models for suicide risk and pharmacogenomic response.

2. Methodology

The MadManic cohort is a large-scale, longitudinal initiative initiated in 2022 in Madrid, Spain, designed to integrate multi-omics data with digital phenotyping.

Study Design and Recruitment

• Participants: The cohort includes over 4,400 participants (~2,300 psychiatric patients and ~2,100 controls) with recruitment ongoing.
• Diagnoses: Patients are recruited from multiple hospitals in the Madrid metropolitan area and include BD, SCZ, MDD, suicide attempters (SA), suicidal ideation (SI), Substance Use Disorders (SUD), and Borderline Personality Disorder (BPD).
• Controls: Healthy controls are recruited from occupational medicine services and nursing schools, screened to exclude any history of psychiatric diagnosis or medication.
• Biospecimens: Over 11,000 biospecimens have been collected, including DNA, RNA, plasma, and serum. Samples are processed within 2–6 hours of extraction at the Centro de Biología Molecular Severo Ochoa (CBMSO).

Data Collection Layers

1. Clinical Phenotyping:
   • Deep phenotyping via structured interviews covering socio-demographics, biometrics, sleep, substance use, disease trajectory (episode counts, polarity), comorbidities, and family history.
   • Suicidality: Assessed using the Columbia–Suicide Severity Rating Scale (C-SSRS) and specific items from the Diagnostic Interview for Genetic Studies (DIGS).
   • Pharmacogenomics: Retrospective assessment of treatment response (e.g., Lithium response via the ALDA scale) and side effects.
2. Digital Phenotyping:
   • Active Monitoring: The MeMind app collects longitudinal self-reported data on mood, sleep, and suicidality via random prompts.
   • Passive Monitoring: The Evidence-based Behaviour (eB2) app utilizes smartphone sensors to track mobility, physical activity, sleep quality, speech prosody, and device usage patterns.
3. Multi-Omics Profiling:
   • Genomics: Genotyping performed using the Axiom Spain Biobank Array (814k markers). Data undergoes rigorous QC, imputation (Haplotype Reference Consortium), and analysis via the Ricopili pipeline.
   • Transcriptomics: RNA sequencing (RNAseq) of blood-derived total RNA using the NovaSeq X Plus platform (~100M paired-end reads per sample).
   • Epigenomics: DNA methylation profiling using the Infinium MethylationEPIC v2.0 BeadChip (>935k CpG sites).

Computational Infrastructure

The study utilizes High-Performance Computing (HPC) resources, including the CCC (UAM), the Drago cluster (CSIC), and the Snellius supercomputer (Netherlands) for collaboration with the Psychiatric Genomics Consortium (PGC). Pipelines include PRS-CSx for risk scoring, CIBERSORTx for immune cell deconvolution, MAJIQ/FRASER for splicing analysis, and methylCIPHER for epigenetic aging.

3. Key Contributions

• Cohort Establishment: Creation of one of the largest Southern European cohorts specifically designed to integrate genomic, transcriptomic, and epigenomic data with digital phenotyping for SMDs.
• Transdiagnostic Framework: A design that allows for the systematic comparison of genetic and molecular signatures across BD, SCZ, and MDD, facilitating the identification of shared vs. disorder-specific mechanisms.
• Digital Integration: Pioneering the combination of longitudinal digital behavioral data (EMA) with multi-omics to improve risk stratification, particularly for suicide.
• Pharmacogenomic Pipeline: Development of specific pipelines to infer metabolizer phenotypes for Cytochrome P450 enzymes and HLA diplotypes to predict drug hypersensitivity and efficacy.
• Population Representation: Addressing the "Northern European bias" in psychiatric genetics by significantly increasing the representation of Southern European ancestry in global meta-analyses.

4. Results (Current Status)

• Sample Processing: Successfully processed >11,000 biospecimens.
• Genotyping: Completed QC and imputation for 5,374 samples (3,526 controls from BNADN + 2,259 MadManic cases), resulting in 7.37 million SNPs. Multidimensional scaling (MDS) confirmed the cohort's European ancestry structure.
• Transcriptomics: RNAseq data generated for 725 individuals (444 BD cases, 281 controls) with high-quality metrics (mean Phred score >39, >50M unique reads).
• Epigenomics: MethylationEPIC v2.0 data generated for 239 individuals (130 BD, 109 controls) with high bisulfite conversion efficiency (95.58%) and successful retention of ~826k probes.
• Data Availability: Individual-level data is restricted; summary statistics are available upon request for collaborative research.

5. Significance

The MadManic cohort represents a critical step toward precision psychiatry. By bridging the gap between static genomic data and dynamic digital phenotyping, it offers a unique resource to:

1. Refine Risk Prediction: Develop machine learning models that combine genetic liability (PRS) with real-time behavioral markers to predict suicide risk and disease relapse more accurately than clinical scales alone.
2. Uncover Biological Mechanisms: Utilize integrative multi-omics (TWAS, ASE, EWAS) to move from statistical associations to functional causal genes and regulatory pathways.
3. Personalize Treatment: Enable pharmacogenomic-guided prescribing to optimize treatment response and minimize adverse effects, particularly for mood stabilizers and antipsychotics.
4. Global Equity: Enhance the diversity of psychiatric genetic databases, ensuring that precision medicine tools are applicable to diverse global populations, not just those of Northern European descent.

The cohort is currently expanding, with future plans to incorporate untargeted metabolomics and proteomics to further close the gap in understanding the molecular landscape of severe mental disorders.