Shamil Sunyaev

Professor

Harvard University
Country flagUnited States
ApplyKite TwinLearns as you apply

Make your next move.

Prepare your application

Review drafts before you send them.

Explore related searches

Articles (9)

Most human variants identified by genome-wide association studies are believed to affect traits by altering gene expression. This belief is supported by considerable circumstantial evidence, but statistical methods are unable to link most trait-associated variants to gene expression—a problem we refer to as “missing regulation.” Many explanations have been proposed, including the possibility that natural selection on gene expression limits power. Here, we take a novel approach to the question of missing regulation, beginning with the observation that the majority of trait-associated variants alter gene expression in two non-human species: cattle and pigs. We explain this discrepancy by comparing the species’ evolutionary histories. The observed differences in regulatory variants are consistent with selection on human gene regulation and increased genetic drift due to agricultural breeding. The differences are not limited to specific genes and reflect increased ascertainment of regulatory variants that are distal to genes. Additionally, we show that trait-associated gene regulation in cattle and pigs matches observed patterns from complex-trait genetics in humans, and may reflect currently unobserved trait-associated regulation in humans.

Year:

2026

Statistical colocalization testing methods can determine if the same single-nucleotide polymorphism (SNP) underlies both a genome-wide association study (GWAS) locus as well as an expression quantitative trait (eQTL) locus. This can nominate potential mechanistic pathways from SNPs to genes to traits, while providing cell type or tissue context. Surprisingly, systematic colocalization testing with bulk-tissue eQTLs fails to link the majority of GWAS loci with gene expression changes. Mapping eQTLs with single-cell expression data has the potential to reveal the missing regulatory effects of GWAS variants. However, current pseudobulk cluster-based approaches may be underpowered when clustering accuracy is imperfect or with an incorrectly selected cluster resolution. To improve power of single-cell colocalization tests, we developed a cluster-free method, scJLIM. By modeling eQTL interactions with continuous cell states (e.g., principal components), scJLIM estimates eQTL significance and colocalization in individual cells. We benchmarked our method with simulated data, demonstrating improvements in power over pseudobulk methods. In our main applications, we used scJLIM to analyze blood and brain scRNA-seq datasets paired with autoimmune and neurological disease GWAS, respectively. We identified nearly twice as many total colocalizations compared with traditional pseudobulk analyses carried out within the major cell populations of these tissues. Aligning with a recent experimental study, we highlighted an example of the ETS2 gene colocalizing with an inflammatory bowel disease GWAS locus in a subset of myeloid cells. For Parkinson’s disease (PD), our results pointed to TRPV2 as a potential gene of interest, corroborated by transcriptional changes in both post-mortem PD brains and iPSC-derived neuronal models of alpha-synucleinopathy.

Year:

2025

Using the Telomere-to-Telomere reference, we assembled the distribution of simple repeat lengths present in the human genome. Analyzing over two hundred mammalian genomes, we found remarkable consistency in the shape of the distribution across evolutionary epochs. All observed genomes harbor an excess of long repeats, which are potentially prone to developing into repeat expansion disorders. We measured mutation rates for repeat length instability, quantitatively modeled the per-generation action of mutations, and observed the corresponding long-term behavior shaping the repeat length distribution. We found that short repetitive sequences appear to be a straightforward consequence of random substitution. Evolving largely independently, longer repeats (above roughly 10 nt) emerge and persist in a rapidly mutating dynamic balance between expansion, contraction and interruption. These mutational processes, collectively, are sufficient to explain the abundance of long repeats, without invoking natural selection. Our analysis constrains properties of molecular mechanisms responsible for maintaining genome fidelity that underlie repeat instability.

Year:

2025

There are two primary approaches to study the genetic basis of human phenotypes. Experiments in model systems generate interpretable gene networks but, in isolation, do not establish relevance to the human condition. Statistical genetics identifies relevant association signals at the variant or gene level but lacks tools to test specific mechanistic models, as existing methods do not incorporate the topology of gene-gene interactions. We bridge these two strategies by introducing a method that competitively tests network hypotheses with rare variant associations. A hierarchical model-based association test NERINE for the first time incorporates gene network topology while remaining resilient to network inaccuracies. We demonstrate NERINE’s ability to test network hypotheses derived from both canonical pathway databases and model system screens. Comprehensive database-wide search of pathway networks with NERINE uncovers compelling associations for breast cancer, cardiovascular diseases, and type II diabetes, which are undetected by single-gene tests. Testing bespoke networks from experimental screens targeting key PD pathologies: dopaminergic neuron survival and α-synuclein pathobiology, NERINE highlights rare variant burden in gene modules related to autophagy, vesicle trafficking, and protein homeostasis. Genome-scale CRISPRi-screening of α-synuclein toxicity modifiers in human neurons and NERINE converge on PRL , revealing an intraneuronal α-synuclein / prolactin stress response that may impact resilience to PD pathologies.

Year:

2025

In renewing tissues, mutations conferring selective advantage may result in clonal expansions 1–4 . In contrast to somatic tissues, mutations driving clonal expansions in spermatogonia (CES) are also transmitted to the next generation. This results in an effective increase of de novo mutation rate for CES drivers 5–8 . CES was originally discovered through extreme recurrence of de novo mutations causing Apert syndrome 5 . Here, we develop a systematic approach to discover CES drivers as hotspots of human de novo mutation. Our analysis of 54,715 trios ascertained for rare conditions 9–13, 6,065 control trios12,14–19, and population variation from 807,162 mostly healthy individuals 20 identifies genes manifesting rates of de novo mutations inconsistent with plausible models of disease ascertainment. We propose 23 genes hypermutable at loss-of-function (LoF) sites as candidate CES drivers. An additional 17 genes feature hypermutable missense mutations at individual positions, suggesting CES acting through gain-of-function (GoF). CES increases the average mutation rate ∼17-fold for LoF genes in both control trios and sperm and ∼500-fold for pooled GoF sites in sperm. Positive selection in the male germline elevates the prevalence of genetic disorders and increases polymorphism levels, masking the effect of negative selection in human populations. Despite the excess of mutations in disease cohorts for 19 LoF CES driver candidates, only 9 show clear evidence of disease causality 21 , suggesting that CES may lead to false-positive disease associations.

Year:

2025

Variation in complex traits is controlled by multiple genes. The prevailing assumption is that such polygenic complex traits are underpinned by variation in elementary molecular traits, such as gene expression, which themselves have a simple genetic basis. Here, we review recent advances that reveal the captivating complexity of gene regulation: the cell type, time point, and magnitude of gene expression are not merely dependent on a couple of regulators; rather, they result from a probabilistic process shaped by cis- and trans-regulatory elements collaboratively integrating internal and external cues with the tightly regulated dynamics of DNA. In addition, the finding that genetic variants linked to complex diseases in humans often do not co-localize with quantitative trait loci modulating gene expression, along with the role of nonfunctional transcription factor (TF) binding sites, suggests that some of the genetic effects influencing gene expression variation may be indirect. If the number of genomic positions responsible for TF binding, TF binding site search time, DNA conformation and accessibility as well as regulation of all trans-acting factors is indeed vast, is it plausible that the complexity of elementary molecular traits approaches the complexity of higher-level organismal traits? Although it is hard to know the answer to this question, we motivate it by reviewing the complexity of the molecular machinery further.

Year:

2024

Results from genome-wide association studies (GWAS) enable inferences about the balance of evolutionary forces maintaining genetic variation underlying common diseases and other genetically complex traits. Natural selection is a major force shaping variation, and understanding it is necessary to explain the genetic architecture and prevalence of heritable diseases. Here, we analyze data for 27 traits, including anthropometric traits, metabolic traits, and binary diseases—both early-onset and post-reproductive. We develop an inference framework to test existing population genetics models based on the joint distribution of allelic effect sizes and frequencies of trait-associated variants. A majority of traits have GWAS results that are inconsistent with neutral evolution or long-term directional selection (selection against a trait or against disease risk). Instead, we find that most traits show consistency with stabilizing selection, which acts to preserve an intermediate trait value or disease risk. Our observations also suggest that selection may reflect pleiotropy, with each variant influenced by associations with multiple selected traits.

Year:

2024

Whether neurodegenerative diseases linked to misfolding of the same protein share genetic risk drivers or whether different protein-aggregation pathologies in neurodegeneration are mechanistically related remains uncertain. Conventional genetic analyses are underpowered to address these questions. Through careful selection of patients based on protein aggregation phenotype (rather than clinical diagnosis) we can increase statistical power to detect associated variants in a targeted set of genes that modify proteotoxicities. Genetic modifiers of alpha-synuclein (ɑS) and beta-amyloid (Aβ) cytotoxicity in yeast are enriched in risk factors for Parkinson’s disease (PD) and Alzheimer’s disease (AD), respectively. Here, along with known AD/PD risk genes, we deeply sequenced exomes of 430 ɑS/Aβ modifier genes in patients across alpha-synucleinopathies (PD, Lewy body dementia and multiple system atrophy). Beyond known PD genes GBA1 and LRRK2 , rare variants AD genes ( CD33 , CR1 and PSEN2 ) and Aβ toxicity modifiers involved in RhoA/actin cytoskeleton regulation ( ARGHEF1, ARHGEF28, MICAL3, PASK, PKN2, PSEN2 ) were shared risk factors across synucleinopathies. Actin pathology occurred in iPSC synucleinopathy models and RhoA downregulation exacerbated ɑS pathology. Even in sporadic PD, the expression of these genes was altered across CNS cell types. Genome-wide CRISPR screens revealed the essentiality of PSEN2 in both human cortical and dopaminergic neurons, and PSEN2 mutation carriers exhibited diffuse brainstem and cortical synucleinopathy independent of AD pathology. PSEN2 contributes to a common-risk signal in PD GWAS and regulates ɑS expression in neurons. Our results identify convergent mechanisms across synucleinopathies, some shared with AD.

Year:

2024

Genomics for rare disease diagnosis has advanced at a rapid pace due to our ability to perform "N-of-1" analyses on individual patients with ultra-rare diseases. The increasing sizes of ultra-rare disease cohorts internationally newly enables cohort-wide analyses for new discoveries, but well-calibrated statistical genetics approaches for jointly analyzing these patients are still under development. The Undiagnosed Diseases Network (UDN) brings multiple clinical, research and experimental centers under the same umbrella across the United States to facilitate and scale N-of-1 analyses. Here, we present the first joint analysis of whole genome sequencing data of UDN patients across the network. We introduce new, well-calibrated statistical methods for prioritizing disease genes with de novo recurrence and compound heterozygosity. We also detect pathways enriched with candidate and known diagnostic genes. Our computational analysis, coupled with a systematic clinical review, recapitulated known diagnoses and revealed new disease associations. We further release a software package, RaMeDiES, enabling automated cross-analysis of deidentified sequenced cohorts for new diagnostic and research discoveries. Gene-level findings and variant-level information across the cohort are available in a public-facing browser (https://dbmi-bgm.github.io/udn-browser/). These results show that N-of-1 efforts should be supplemented by a joint genomic analysis across cohorts.

Year:

2024

Collaborators (12)

Raheleh Rahbari

Group Leader

Wellcome Sanger Institute

UNITED KINGDOM
View details

Julian Martinez-Agosto

Associate Professor (with tenure)

David Geffen School of Medicine at UCLA

UNITED STATES
View details

Isaac Kohane

Professor

Harvard University

UNITED STATES
View details

Kelvin Luk

Associate Professor

University of Pennsylvania

UNITED STATES
View details

Benjamin Neale

Associate Professor

Broad Institute

UNITED STATES
View details

Vladimir Seplyarskiy

Assistant Professor

University of Texas Southwestern Medical Center

UNITED STATES
View details

Juliette de Meaux

University of Cologne

GERMANY
View details

Evan Koch

Assistant Professor

Yale University

UNITED STATES
View details

Sergei Mirkin

Endowed Chair Professor

Tufts University

UNITED STATES
View details

Chris Cotsapas

Professor

Yale University School of Medicine

UNITED STATES
View details

Nathan Stitziel

Professor

Washington University in Saint Louis School of Medicine

UNITED STATES
View details

Stanley Nelson

Professor

David Geffen School of Medicine at UCLA

UNITED STATES
View details