Jia Wen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Jia Wen has a total of 293 co-authors. Here are the top co-authors by number of shared publications:
- Yun Li
Department of Biostatistics, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA
8 shared publications - Jin P. Szatkiewicz
Department of Genetics, University of North Carolina at Chapel Hill, Chapel Hill, USA
6 shared publications - Patrick F. Sullivan
Department of Genetics, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA
6 shared publications - Quan Sun
Department of Biostatistics, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA
6 shared publications - Alexander P. Reiner
University of Washington
5 shared publications - Laura M. Raffield
University of North Carolina at Chapel Hill
4 shared publications
Latest Publications
- Open Access
- 30 September 2026
- Published by Wolters Kluwer Health in Circulation: Genomic and Precision Medicine
Polygenic Risk Scores of Coronary Heart Disease in Chronic Kidney Disease Population: The CRIC Study
BACKGROUND: Cardiovascular disease is a major cause of death in individuals with chronic kidney disease; thus, improving its prediction is critical for cardiovascular disease risk reduction. Studies have not assessed whether a polygenic risk score (PRS) of coronary heart disease (CHD) developed from genetic studies in the general population can improve cardiovascular disease risk prediction in individuals with chronic kidney disease, especially when PRS is added to the American Heart Association Predicting Risk of Cardiovascular Disease EVENTs (PREVENT) equations. METHODS: Focusing on a 10-year composite CHD event, we applied PRSs of CHD obtained from the polygenic score catalog to the multi-center CRIC (Chronic Renal Insufficiency Cohort) participants, selected the best-performing PRS and evaluated the added value of the best PRS while accounting for PREVENT linear predictors. Three Cox proportional hazard models were fitted with PRS only, PREVENT predictors only, and PRS+PREVENT risk factors. We evaluated the prediction accuracy, discrimination, and calibration of the combined model compared with the PREVENT model. RESULTS: A PRS of CHD was significantly associated with a 21% hazard of CHD in the PRS+PREVENT risk model (95% CI, [1.06–1.39], 2-sided Wald test P =5.5×10 −3 ). The PRS+PREVENT model demonstrated a higher area under the curve (72.4% [95% CI, 0.66–0.79] versus 70.8% [95% CI, 0.64–0.77]), better discrimination ability, and improved calibration compared with the PREVENT model predictors. Individuals were more correctly classified in prediction quartiles using the PRS+PREVENT model for the 10-year CHD event proportions. CONCLUSIONS: Our results demonstrate the added value of a CHD PRS to the PREVENT risk prediction models in individuals with chronic kidney disease. The combined PRS+PREVENT model has the potential to improve clinical decisions on disease prevention and treatment strategies.
- 30 September 2026
- Published by Wolters Kluwer Health in Circulation: Genomic and Precision Medicine
Open AccessPolygenic Risk Scores of Coronary Heart Disease in Chronic Kidney Disease Population: The CRIC Study
BACKGROUND: Cardiovascular disease is a major cause of death in individuals with chronic kidney disease; thus, improving its prediction is critical for cardiovascular disease risk reduction. Studies have not assessed whether a polygenic risk score (PRS) of coronary heart disease (CHD) developed from genetic studies in the general population can improve cardiovascular disease risk prediction in individuals with chronic kidney disease, especially when PRS is added to the American Heart Association Predicting Risk of Cardiovascular Disease EVENTs (PREVENT) equations. METHODS: Focusing on a 10-year composite CHD event, we applied PRSs of CHD obtained from the polygenic score catalog to the multi-center CRIC (Chronic Renal Insufficiency Cohort) participants, selected the best-performing PRS and evaluated the added value of the best PRS while accounting for PREVENT linear predictors. Three Cox proportional hazard models were fitted with PRS only, PREVENT predictors only, and PRS+PREVENT risk factors. We evaluated the prediction accuracy, discrimination, and calibration of the combined model compared with the PREVENT model. RESULTS: A PRS of CHD was significantly associated with a 21% hazard of CHD in the PRS+PREVENT risk model (95% CI, [1.06–1.39], 2-sided Wald test P =5.5×10 −3 ). The PRS+PREVENT model demonstrated a higher area under the curve (72.4% [95% CI, 0.66–0.79] versus 70.8% [95% CI, 0.64–0.77]), better discrimination ability, and improved calibration compared with the PREVENT model predictors. Individuals were more correctly classified in prediction quartiles using the PRS+PREVENT model for the 10-year CHD event proportions. CONCLUSIONS: Our results demonstrate the added value of a CHD PRS to the PREVENT risk prediction models in individuals with chronic kidney disease. The combined PRS+PREVENT model has the potential to improve clinical decisions on disease prevention and treatment strategies.
- Top Cited PapersOpen Access
- 22 January 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Cross-ancestry comparison of aptamer and antibody protein measures
Measures from affinity-proteomics platforms often correlate poorly, challenging interpretation of protein associations with genetic variants and phenotypes. Here, we examine 2157 proteins measured on both SomaScan 7k and Olink Explore 3072 across 1930 participants with genetic similarity to European, African, East Asian, and Admixed American ancestry references. Inter-platform correlation coefficients for these 2157 proteins follow a bimodal distribution (median r = 0.30). We evaluate protein measure associations with genetic variants, and find approximately 25-30% of the signals on each platform are likely driven by protein-altering variants. We highlight 80 proteins that correlate differently across ancestry groups likely in part due to differing protein-altering variant frequencies by ancestry. Furthermore, adjustment for protein-altering variants with opposite directions of effect by platform improves inter-platform protein measure correlation and results in more concordant genetic and phenotypic associations. Hence, protein-altering variants need to be accounted for across ancestries to facilitate platform-concordant and accurate protein measurement.
- 22 January 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Top Cited PapersOpen AccessCross-ancestry comparison of aptamer and antibody protein measures
Measures from affinity-proteomics platforms often correlate poorly, challenging interpretation of protein associations with genetic variants and phenotypes. Here, we examine 2157 proteins measured on both SomaScan 7k and Olink Explore 3072 across 1930 participants with genetic similarity to European, African, East Asian, and Admixed American ancestry references. Inter-platform correlation coefficients for these 2157 proteins follow a bimodal distribution (median r = 0.30). We evaluate protein measure associations with genetic variants, and find approximately 25-30% of the signals on each platform are likely driven by protein-altering variants. We highlight 80 proteins that correlate differently across ancestry groups likely in part due to differing protein-altering variant frequencies by ancestry. Furthermore, adjustment for protein-altering variants with opposite directions of effect by platform improves inter-platform protein measure correlation and results in more concordant genetic and phenotypic associations. Hence, protein-altering variants need to be accounted for across ancestries to facilitate platform-concordant and accurate protein measurement.
- Open Access
- 22 November 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1)
Integrating whole genome and transcriptome sequencing to characterize the genetic architecture of isoform variation
We present a whole-blood isoform ratio QTL (irQTL) resource by analyzing genome-wide isoform-to-gene expression ratios using sequencing data. In Framingham Heart Study (FHS, n = 2622) discovery, we identify over 1.1 million cis -irQTLs (minor allele frequency [MAF] ≥ 0.01, ±1 Mb of 10,883 isoform transcripts, P < 5 × 10 −8 ) across 4,971 genes. Among 11,425 sentinel cis -irQTLs, 72% replicate ( P < 1 × 10 −4 ) in the Women’s Health Initiative (WHI; n = 2005). Notably, 20% of cis -irQTLs have no significant association with overall gene expression, indicating isoform-specific regulation. These variants are enriched at splice donor/acceptor sites and genome-wide association study loci ( P < 1 × 10 −10 ). We also identify 1870 sentinel trans -irQTLs (MAF ≥ 0.01, P < 1.5 × 10 −13 ) for 1,084 isoforms across 590 genes, and 2327 rare cis -irQTLs (0.003 < MAF < 0.01) for 2467 isoforms of 1428 genes in FHS, with external replication rates of 61% and 41% in WHI, respectively. We highlight rs12898397 in ULK3 , which alters splice site usage and reduces expression of a full-length isoform. Mendelian randomization supports a causal role between this isoform shift and reduced diastolic blood pressure. These findings highlight the power of irQTL mapping to uncover transcript-specific regulatory mechanisms underlying complex traits.
- 22 November 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1)
Open AccessIntegrating whole genome and transcriptome sequencing to characterize the genetic architecture of isoform variation
We present a whole-blood isoform ratio QTL (irQTL) resource by analyzing genome-wide isoform-to-gene expression ratios using sequencing data. In Framingham Heart Study (FHS, n = 2622) discovery, we identify over 1.1 million cis -irQTLs (minor allele frequency [MAF] ≥ 0.01, ±1 Mb of 10,883 isoform transcripts, P < 5 × 10 −8 ) across 4,971 genes. Among 11,425 sentinel cis -irQTLs, 72% replicate ( P < 1 × 10 −4 ) in the Women’s Health Initiative (WHI; n = 2005). Notably, 20% of cis -irQTLs have no significant association with overall gene expression, indicating isoform-specific regulation. These variants are enriched at splice donor/acceptor sites and genome-wide association study loci ( P < 1 × 10 −10 ). We also identify 1870 sentinel trans -irQTLs (MAF ≥ 0.01, P < 1.5 × 10 −13 ) for 1,084 isoforms across 590 genes, and 2327 rare cis -irQTLs (0.003 < MAF < 0.01) for 2467 isoforms of 1428 genes in FHS, with external replication rates of 61% and 41% in WHI, respectively. We highlight rs12898397 in ULK3 , which alters splice site usage and reduces expression of a full-length isoform. Mendelian randomization supports a causal role between this isoform shift and reduced diastolic blood pressure. These findings highlight the power of irQTL mapping to uncover transcript-specific regulatory mechanisms underlying complex traits.
- Open Access
- 13 February 2025
- Published by Springer Nature in Research Square
Cross-Ancestry Comparison of Aptamer and Antibody Proteomics Measures
Measures from affinity-proteomics platforms often correlate poorly, challenging interpretation of protein associations with genetic variants (pQTL) and phenotypes. Here, we examined 2,157 proteins measured on both SomaScan 7k and Olink Explore 3072 across 1,930 participants with genetic similarity to European, African, East Asian, and Admixed American ancestry references. Inter-platform correlation coefficients for these 2,157 proteins followed a bimodal distribution (median r=0.30). Protein measures were associated with variants (pQTLs), and one-third of the pQTL signals were driven by protein-altering variants (PAVs). We highlight 80 proteins that correlate differently across ancestry groups likely due to differing PAV frequencies by ancestry. Furthermore, adjustment for PAVs with opposite directions of effect by platform improved inter-platform protein measure correlation and resulted in more concordant genetic and phenotypic associations. Hence, PAVs need to be accounted for across ancestries to facilitate platform-concordant and accurate protein measurement.
- 13 February 2025
- Published by Springer Nature in Research Square
Open AccessCross-Ancestry Comparison of Aptamer and Antibody Proteomics Measures
Measures from affinity-proteomics platforms often correlate poorly, challenging interpretation of protein associations with genetic variants (pQTL) and phenotypes. Here, we examined 2,157 proteins measured on both SomaScan 7k and Olink Explore 3072 across 1,930 participants with genetic similarity to European, African, East Asian, and Admixed American ancestry references. Inter-platform correlation coefficients for these 2,157 proteins followed a bimodal distribution (median r=0.30). Protein measures were associated with variants (pQTLs), and one-third of the pQTL signals were driven by protein-altering variants (PAVs). We highlight 80 proteins that correlate differently across ancestry groups likely due to differing PAV frequencies by ancestry. Furthermore, adjustment for PAVs with opposite directions of effect by platform improved inter-platform protein measure correlation and resulted in more concordant genetic and phenotypic associations. Hence, PAVs need to be accounted for across ancestries to facilitate platform-concordant and accurate protein measurement.
- Open Access
- 12 November 2024
- Published by Springer Nature in Molecular Psychiatry
- Vol. 30 (5) , 2009-2016
Genome-wide copy number variation association study in anorexia nervosa
This study represents the first large-scale investigation of rare (CD36 (7q21.11) gene, which transports long-chain fatty acids into cells. CNVRs intersecting genes previously related to neurodevelopmental traits include deletions of NRXN1 intron 5 (2p16.3), IMMP2L (7q31.1), and PTPRD (9p23). Overall, given that our study is well powered to detect the CNV burden level reported for schizophrenia, we can conclude that rare CNVs have a limited role in the etiology of AN, as reported for bipolar disorder. Our nominal associations for the 21 discovered CNVRs are consistent with AN being a metabo-psychiatric trait, as demonstrated by the common genetic architecture of AN, and we provide association results to allow for replication in future research.
- 12 November 2024
- Published by Springer Nature in Molecular Psychiatry
- Vol. 30 (5) , 2009-2016
Open AccessGenome-wide copy number variation association study in anorexia nervosa
This study represents the first large-scale investigation of rare (CD36 (7q21.11) gene, which transports long-chain fatty acids into cells. CNVRs intersecting genes previously related to neurodevelopmental traits include deletions of NRXN1 intron 5 (2p16.3), IMMP2L (7q31.1), and PTPRD (9p23). Overall, given that our study is well powered to detect the CNV burden level reported for schizophrenia, we can conclude that rare CNVs have a limited role in the etiology of AN, as reported for bipolar disorder. Our nominal associations for the 21 discovered CNVRs are consistent with AN being a metabo-psychiatric trait, as demonstrated by the common genetic architecture of AN, and we provide association results to allow for replication in future research.
- 21 September 2023
- Published by Springer Nature in Nature Genetics
- Vol. 55 (10) , 1735-1744
Functional characterization of Alzheimer’s disease genetic variants in microglia
Candidate cis-regulatory elements (cCREs) in microglia demonstrate the most substantial enrichment for Alzheimer’s disease (AD) heritability compared to other brain cell types. However, whether and how these genome-wide association studies (GWAS) variants contribute to AD remain elusive. Here we prioritize 308 previously unreported AD risk variants at 181 cCREs by integrating genetic information with microglia-specific 3D epigenome annotation. We further establish the link between functional variants and target genes by single-cell CRISPRi screening in microglia. In addition, we show that AD variants exhibit allelic imbalance on target gene expression. In particular, rs7922621 is the effective variant in controlling TSPAN14 expression among other nominated variants in the same cCRE and exerts multiple physiological effects including reduced cell surface ADAM10 and altered soluble TREM2 (sTREM2) shedding. Our work represents a systematic approach to prioritize and characterize AD-associated variants and provides a roadmap for advancing genetic association to experimentally validated cell-type-specific phenotypes and mechanisms.
- 21 September 2023
- Published by Springer Nature in Nature Genetics
- Vol. 55 (10) , 1735-1744
Functional characterization of Alzheimer’s disease genetic variants in microglia
Candidate cis-regulatory elements (cCREs) in microglia demonstrate the most substantial enrichment for Alzheimer’s disease (AD) heritability compared to other brain cell types. However, whether and how these genome-wide association studies (GWAS) variants contribute to AD remain elusive. Here we prioritize 308 previously unreported AD risk variants at 181 cCREs by integrating genetic information with microglia-specific 3D epigenome annotation. We further establish the link between functional variants and target genes by single-cell CRISPRi screening in microglia. In addition, we show that AD variants exhibit allelic imbalance on target gene expression. In particular, rs7922621 is the effective variant in controlling TSPAN14 expression among other nominated variants in the same cCRE and exerts multiple physiological effects including reduced cell surface ADAM10 and altered soluble TREM2 (sTREM2) shedding. Our work represents a systematic approach to prioritize and characterize AD-associated variants and provides a roadmap for advancing genetic association to experimentally validated cell-type-specific phenotypes and mechanisms.
- Open Access
- 14 February 2023
- Published by Springer Nature in Research Square
DeepGWAS: Enhance GWAS Signals for Neuropsychiatric Disorders via Deep Neural Network
Genetic dissection of neuropsychiatric disorders can potentially reveal novel therapeutic targets. While genome-wide association studies (GWAS) have tremendously advanced our understanding, we approach a sample size bottleneck (i.e., the number of cases needed to identify >90% of all loci is impractical). Therefore, computationally enhancing GWAS on existing samples may be particularly valuable. Here, we describe DeepGWAS, a deep neural network-based method to enhance GWAS by integrating GWAS results with linkage disequilibrium and brain-related functional annotations. DeepGWAS enhanced schizophrenia (SCZ) loci by ~3X when applied to the largest European GWAS, and 21.3% enhanced loci were validated by the latest multi-ancestry GWAS. Importantly, DeepGWAS models can be transferred to other neuropsychiatric disorders. Transferring SCZ-trained models to Alzheimer’s disease and major depressive disorder, we observed 1.3-17.6X detected loci compared to standard GWAS, among which 27-40% were validated by other GWAS studies. We anticipate DeepGWAS to be a powerful tool in GWAS studies.
- 14 February 2023
- Published by Springer Nature in Research Square
Open AccessDeepGWAS: Enhance GWAS Signals for Neuropsychiatric Disorders via Deep Neural Network
Genetic dissection of neuropsychiatric disorders can potentially reveal novel therapeutic targets. While genome-wide association studies (GWAS) have tremendously advanced our understanding, we approach a sample size bottleneck (i.e., the number of cases needed to identify >90% of all loci is impractical). Therefore, computationally enhancing GWAS on existing samples may be particularly valuable. Here, we describe DeepGWAS, a deep neural network-based method to enhance GWAS by integrating GWAS results with linkage disequilibrium and brain-related functional annotations. DeepGWAS enhanced schizophrenia (SCZ) loci by ~3X when applied to the largest European GWAS, and 21.3% enhanced loci were validated by the latest multi-ancestry GWAS. Importantly, DeepGWAS models can be transferred to other neuropsychiatric disorders. Transferring SCZ-trained models to Alzheimer’s disease and major depressive disorder, we observed 1.3-17.6X detected loci compared to standard GWAS, among which 27-40% were validated by other GWAS studies. We anticipate DeepGWAS to be a powerful tool in GWAS studies.
- Open Access
- 8 December 2022
- Published by Springer Nature in Nature Communications
- Vol. 13 (1) , 1-18
Whole genome sequencing identifies structural variants contributing to hematologic traits in the NHLBI TOPMed program
Genome-wide association studies have identified thousands of single nucleotide variants and small indels that contribute to variation in hematologic traits. While structural variants are known to cause rare blood or hematopoietic disorders, the genome-wide contribution of structural variants to quantitative blood cell trait variation is unknown. Here we utilized whole genome sequencing data in ancestrally diverse participants of the NHLBI Trans Omics for Precision Medicine program (N = 50,675) to detect structural variants associated with hematologic traits. Using single variant tests, we assessed the association of common and rare structural variants with red cell-, white cell-, and platelet-related quantitative traits and observed 21 independent signals (12 common and 9 rare) reaching genome-wide significance. The majority of these associations (N = 18) replicated in independent datasets. In genome-editing experiments, we provide evidence that a deletion associated with lower monocyte counts leads to disruption of an S1PR3 monocyte enhancer and decreased S1PR3 expression.
- 8 December 2022
- Published by Springer Nature in Nature Communications
- Vol. 13 (1) , 1-18
Open AccessWhole genome sequencing identifies structural variants contributing to hematologic traits in the NHLBI TOPMed program
Genome-wide association studies have identified thousands of single nucleotide variants and small indels that contribute to variation in hematologic traits. While structural variants are known to cause rare blood or hematopoietic disorders, the genome-wide contribution of structural variants to quantitative blood cell trait variation is unknown. Here we utilized whole genome sequencing data in ancestrally diverse participants of the NHLBI Trans Omics for Precision Medicine program (N = 50,675) to detect structural variants associated with hematologic traits. Using single variant tests, we assessed the association of common and rare structural variants with red cell-, white cell-, and platelet-related quantitative traits and observed 21 independent signals (12 common and 9 rare) reaching genome-wide significance. The majority of these associations (N = 18) replicated in independent datasets. In genome-editing experiments, we provide evidence that a deletion associated with lower monocyte counts leads to disruption of an S1PR3 monocyte enhancer and decreased S1PR3 expression.
- Open Access
Increased Prevalence of Rare Copy Number Variants in Treatment-Resistant Psychosis
Background: It remains unknown why ∼30% of patients with psychotic disorders fail to respond to treatment. Previous genomic investigations into treatment-resistant psychosis have been inconclusive, but some evidence suggests a possible link between rare disease-associated copy number variants (CNVs) and worse clinical outcomes in schizophrenia. Here, we test whether schizophrenia-associated CNVs are more prevalent in patients with treatment-resistant psychotic symptoms compared to previously published schizophrenia cases not selected for treatment-resistance. Methods: CNVs were identified using chromosomal microarrays and exome sequencing in 509 patients with treatment-resistant psychosis (a lack of clinical response to ≥ 3 adequate antipsychotic medication trials over at least five years of psychiatric hospitalization). Prevalence of schizophrenia-associated CNVs in this sample was compared against a previous large schizophrenia cohort study. Results: In total, 47 cases (9.2%) carried at least one CNV with known or possible neuropsychiatric risk. The prevalence of schizophrenia-associated CNVs (n=21; 4.1%) was significantly increased compared to a previous schizophrenia cohort study ( p = 0.005322; OR = 1.93). This increase in prevalence was primarily due to duplications at 15q11.2-q13.1 and 16p11.2, which were independently associated with treatment-resistance in pairwise loci-based analysis. Conclusions: These findings suggest that rare schizophrenia-associated CNVs, particularly duplications of 15q11.2-q13.1 and 16p11.2, may serve as biological entry points for studying treatment resistance. Further investigation will be necessary to elucidate the spectrum of phenotypic characteristics observed in adult psychiatric patients with disease-associated CNVs.
Open AccessIncreased Prevalence of Rare Copy Number Variants in Treatment-Resistant Psychosis
Background: It remains unknown why ∼30% of patients with psychotic disorders fail to respond to treatment. Previous genomic investigations into treatment-resistant psychosis have been inconclusive, but some evidence suggests a possible link between rare disease-associated copy number variants (CNVs) and worse clinical outcomes in schizophrenia. Here, we test whether schizophrenia-associated CNVs are more prevalent in patients with treatment-resistant psychotic symptoms compared to previously published schizophrenia cases not selected for treatment-resistance. Methods: CNVs were identified using chromosomal microarrays and exome sequencing in 509 patients with treatment-resistant psychosis (a lack of clinical response to ≥ 3 adequate antipsychotic medication trials over at least five years of psychiatric hospitalization). Prevalence of schizophrenia-associated CNVs in this sample was compared against a previous large schizophrenia cohort study. Results: In total, 47 cases (9.2%) carried at least one CNV with known or possible neuropsychiatric risk. The prevalence of schizophrenia-associated CNVs (n=21; 4.1%) was significantly increased compared to a previous schizophrenia cohort study ( p = 0.005322; OR = 1.93). This increase in prevalence was primarily due to duplications at 15q11.2-q13.1 and 16p11.2, which were independently associated with treatment-resistance in pairwise loci-based analysis. Conclusions: These findings suggest that rare schizophrenia-associated CNVs, particularly duplications of 15q11.2-q13.1 and 16p11.2, may serve as biological entry points for studying treatment resistance. Further investigation will be necessary to elucidate the spectrum of phenotypic characteristics observed in adult psychiatric patients with disease-associated CNVs.
- Open Access
- 8 March 2022
- Published by Public Library of Science (PLoS) in PLOS Genetics
- Vol. 18 (3) , e1010102
THUNDER: A reference-free deconvolution method to infer cell type proportions from bulk Hi-C data
Hi-C data provide population averaged estimates of three-dimensional chromatin contacts across cell types and states in bulk samples. Effective analysis of Hi-C data entails controlling for the potential confounding factor of differential cell type proportions across heterogeneous bulk samples. We propose a novel unsupervised deconvolution method for inferring cell type composition from bulk Hi-C data, the Two-step Hi-c UNsupervised DEconvolution appRoach (THUNDER). We conducted extensive simulations to test THUNDER based on combining two published single-cell Hi-C (scHi-C) datasets. THUNDER more accurately estimates the underlying cell type proportions compared to reference-free methods (e.g., TOAST, and NMF) and is more robust than reference-dependent methods (e.g. MuSiC). We further demonstrate the practical utility of THUNDER to estimate cell type proportions and identify cell-type-specific interactions in Hi-C data from adult human cortex tissue samples. THUNDER will be a useful tool in adjusting for varying cell type composition in population samples, facilitating valid and more powerful downstream analysis such as differential chromatin organization studies. Additionally, THUNDER estimated contact profiles provide a useful exploratory framework to investigate cell-type-specificity of the chromatin interactome while experimental data is still rare.
- 8 March 2022
- Published by Public Library of Science (PLoS) in PLOS Genetics
- Vol. 18 (3) , e1010102
Open AccessTHUNDER: A reference-free deconvolution method to infer cell type proportions from bulk Hi-C data
Hi-C data provide population averaged estimates of three-dimensional chromatin contacts across cell types and states in bulk samples. Effective analysis of Hi-C data entails controlling for the potential confounding factor of differential cell type proportions across heterogeneous bulk samples. We propose a novel unsupervised deconvolution method for inferring cell type composition from bulk Hi-C data, the Two-step Hi-c UNsupervised DEconvolution appRoach (THUNDER). We conducted extensive simulations to test THUNDER based on combining two published single-cell Hi-C (scHi-C) datasets. THUNDER more accurately estimates the underlying cell type proportions compared to reference-free methods (e.g., TOAST, and NMF) and is more robust than reference-dependent methods (e.g. MuSiC). We further demonstrate the practical utility of THUNDER to estimate cell type proportions and identify cell-type-specific interactions in Hi-C data from adult human cortex tissue samples. THUNDER will be a useful tool in adjusting for varying cell type composition in population samples, facilitating valid and more powerful downstream analysis such as differential chromatin organization studies. Additionally, THUNDER estimated contact profiles provide a useful exploratory framework to investigate cell-type-specificity of the chromatin interactome while experimental data is still rare.