EctoTrace Diagnostics

High Heritability of Schizophrenia

Schizophrenia is a severe psychiatric illness with a significant genetic component. Family¹,², twin, and adoption studies consistently find high heritability, with twin studies providing estimates approaching 80%³. This places schizophrenia among the most heritable complex brain disorders and underscores the importance of genetic factors in its etiology⁴.

However, since even identical twins are not 100% concordant, non-genetic factors must also play a role. This suggests that environmental exposures likely modify an underlying genetic predisposition, a process often mediated by epigenetics.

A Complex, Polygenic Architecture

Despite its high heritability, schizophrenia does not follow a simple Mendelian inheritance model. Instead, its inheritance pattern is consistent with a multifactorial etiology.

Large-scale genome-wide association studies (GWAS) have established that its genetic architecture is highly polygenic, with risk conferred by the cumulative effect of thousands of common variants, each with a small effect size, primarily affecting neurodevelopmental and synaptic processes⁵.

Genetic architecture of complex traits, showing schizophrenia variants. — Genetic risk for schizophrenia is composed of rare, high-penetrance variants (left) and common, low-effect variants (right). Adapted from McCarthy et al. 2008⁶

A Disease Signature in the Brain's Epigenome

Recent analyses of post-mortem brain tissue indicate that schizophrenia is associated with a specific, reproducible epigenetic signature⁷⁻¹⁰. Widespread changes in DNA methylation have been identified in key neural pathways, providing a quantifiable biological signal of the disease state.

This molecular fingerprint, consistently found across multiple cohorts, suggests that while genetic risk is complex, its manifestation in the brain's epigenome may be a more focused and detectable target. The critical challenge is how to access this signal in living individuals.

Manhattan plot showing differential DNA methylation across the human genome in schizophrenia cases vs. controls. — Differential DNA methylation in SCZ cases vs. controls across the genome. Adapted from Hannon et al. 2021⁸

A Historical Precedent & A Modern Approach

The idea of using skin as a proxy for brain development is not new. For decades, researchers noted subtle differences in the dermatoglyphics (fingerprints) of individuals with schizophrenia, viewing them as a physical trace of early neurodevelopmental events that occur in the same embryonic window¹¹,¹².

While this was a pioneering concept, we can now move beyond morphology. With modern molecular tools at single-cell resolution, we aim to detect the epigenetic residue of these same early events, transitioning from a physical trace to a precise molecular signal.

Comparison of fingerprints showing dissociation in a whorl pattern, as studied in schizophrenia research. — Adapted from Raphael & Raphael 1962¹¹

Rationale and Feasibility of the Ectodermal Approach

Our hypothesis is that early developmental insults contributing to schizophrenia risk may leave durable epigenetic marks in both skin and brain. To test the feasibility of detecting these marks, we analyzed public data from the NIH Roadmap Epigenomics Project.

Our preliminary results indicate that key regulatory elements associated with developmental gene regulation exhibit greater conservation between brain and skin than between brain and blood, providing a strong empirical basis for our approach.

Heatmap showing conservation of ChromHMM states across different tissues. — Heatmap showing greater conservation of developmental regulatory elements (e.g., TssA, TssBiv) between brain and skin tissues compared to blood.

Current Tools Are Insufficient

Most bioinformatics tools, like edgeR and DESeq2, analyze each genomic region independently. This "one-peak-at-a-time" approach ignores that regulatory elements work in coordinated networks.

They also inherit statistical assumptions from RNA-seq that are often false for epigenetics, potentially causing real biological signals to be misinterpreted as technical noise. Finally, they fail to integrate prior biological knowledge (e.g., conservation, 3D structure) into the core analysis.

We need tools built specifically for chromatin data that model regulation as it actually works: in context and as a network.

Our Solution: The ChromaStat Framework

To overcome this challenge, we are developing ChromaStat, a computational framework designed to identify conserved epigenetic signals across different tissues. By integrating massive public reference epigenomes with a novel weighted Bayesian model, ChromaStat can:

Model tissue-specific variation to filter out cellular "noise."
Identify conserved regulatory elements active in both skin and brain.
Amplify the shared disease signal that would otherwise be obscured.

This framework is designed to bridge the critical analytical gap, enabling the use of skin as an accessible proxy to detect brain-relevant molecular markers.

Addressing Clinical Heterogeneity

Beyond cross-tissue analysis, the ChromaStat framework is designed to address schizophrenia's vast clinical heterogeneity. By integrating deep molecular data with rich clinical metadata, our platform moves beyond treating patients as a single group. We employ unsupervised machine learning to discover patient subgroups based on shared molecular signatures, and our Bayesian engine can directly model covariates like medication response and age of onset. The ultimate objective is to identify robust biomarkers for patient stratification, paving the way for a personalized medicine approach to guide treatment.

A Dynamic Window into Neurodevelopment

Our vision extends to using this system as a dynamic window into the disease's origins. We can leverage the continuous differentiation of melanocyte stem cells within patient skin as an in situ model of development.

This offers a powerful alternative to in vitro models like iPSCs, which require an artificial process of epigenetic resetting. By studying cells in their native environment, we can directly observe how a patient’s inherent epigenetic landscape affects cell fate, providing unprecedented insight into the functional consequences of schizophrenia's developmental origins.

Our Goals and Future Directions

We are in the final stages of developing the ChromaStat computational framework, with a target completion date in mid-2026. The next step will be to validate the model using single-cell open chromatin data from skin biopsies, testing our central hypothesis that brain-relevant epigenetic patterns can be detected in accessible peripheral tissues.

An IRB protocol has already been drafted and is ready for submission at the start of PGY-1. While this project could form the basis of a K award in a research-focused residency, it is designed to remain adaptable. All wet-lab experiments can be carried out in collaboration with established partners at our affiliated institution, allowing for full clinical engagement during residency. In such scenario, we have also prepared an SBIR application to support this work.

References

Rasic D, et al. Schizophrenia Bulletin. 2014.
Wu EQ, et al. Psychological Medicine. 2006.
Sullivan PF, et al. Archives of General Psychiatry. 2003.
Trotta A, et al. Psychological Medicine. 2015.
Sullivan PF, et al. Nat Rev Genet. 2012.
McCarthy MI, et al. Nat Rev Genet. 2008.
Chen J, et al. JAMA Psychiatry. 2020.
Hannon E, et al. eLife. 2021.
Wockner LF, et al. Transl Psychiatry. 2014.
Girdhar K, et al. Nat Neurosci. 2022.
Raphael T, Raphael LG. JAMA. 1962.
Salvador R, et al. Schizophrenia Bulletin. 2023.