3D genomic features across >50 diverse cell types reveal insights into the genomic architecture of childhood obesity

Data and resource

Datasets used in prior studies are listed in eTable 1. ATAC-seq, RNA-seq, Hi-C, and Capture-C library generation for each cell type is provided in their original published study and their pre-processing pipelines and tools can be found in eMethods.

Definition of cis-Regulatory Elements (cREs)

We intersected ATAC-seq open chromatin regions (OCRs) of each cell type with chromatin conformation capture data determined by Hi-C/Capture-C of the same cell type, and with promoters (-1,500/+500bp of TSS, which were referenced by GENCODE v30.

Childhood obesity GWAS summary statistics

Data on childhood obesity from the EGG consortium was downloaded from www.egg-consortium.org. We used 8,566,179 European ancestry variants (consisting of 8,613 cases and 12,696 controls in stage I; of 921 cases and 1,930 controls in stage II), representing ∼55% of the total 15,504,218 variants observed across all ancestries in the original study³⁵. The sumstats file was reformatted by munge_sumstats.py to standardize with the weighted variants from HapMap v3 within the LDSC baseline, which reduced the variants to 1,217,311 (7.8% of total).

Cell type specific partitioned heritability

We used LDSC (http://www.github.com/bulik/ldsc) v.1.0.1 with --h2 flag to estimate SNP-based heritability of childhood obesity within 4 defined sets of input genomic regions: (1) OCRs, (2) OCRs at gene promoters, (3) cREs, and (4) cREs with an expanded window of ±500 bp. The baseline model LD scores, plink filesets, allele frequencies and variants weights files for the European 1000 genomes project phase 3 in hg38 were downloaded from the provided link (https://alkesgroup.broadinstitute.org/LDSCORE/GRCh38/). The cREs of each cell type were used to create the annotation, which in turn were used to compute annotation-specific LD scores for each cell types cREs set.

Genetic loci included in variant-to-genes mapping

19 sentinel signals that achieved genome-wide significance in the trans-ancestral meta-analysis study³⁵ were leveraged for our analyses. Proxies for each sentinel SNP were queried using TopLD³⁶ and LDlinkR tool³⁷ with the GRCh38 Genome assembly, 1000 Genomes phase 3 v5 variant set, European population, and LD threshold of r²>0.8, which resulted in 771 proxies, including the 21 SNPs from the 99% credible set of the original study (eTable 2).

GWAS-eQTL colocalization

The summary statistics for the European ancestry subset from the EGG consortium GWAS for childhood obesity was used. Common variants (MAF□≥□0.01) from the 1000 Genomes Project v3 samples were used as a reference panel. We used non-overlapped genomic windows of ±250,000 bases extended in both directions from the median genomic position of each of 19 sentinel loci as input. We used ColocQuiaL³⁸ to test genome-wide colocalization of all possible variants included in each inputted window against GTEx v.8 eQTLs associations for all 49 tissues available from https://www.gtexportal.org/home/datasets. A conditional posterior probability of colocalization of 0.8 or greater was imposed.

Enrichment assessment of childhood obesity variants across cell types

To explore the enrichment of childhood obesity GWAS variants across cell types, we carried out Partitioned Linkage Disequilibrium Score Regression (LDSR)³⁹ on all ATAC-seq-defined OCRs for each cell type. We assessed cell-type specific enrichment of GWAS signals in four main categories of genomic regions (Fig. 1A): (1) Total OCRs: open chromatin regions defined by ATAC-seq; (2) Promoter OCRs: the subset of OCRs overlapping a gene promoter; (3) cREs: the subset of OCRs that form chromatin loops (as determined by Hi-C/Promoter Capture-C) with a gene promoter, and are therefore considered putative enhancers or suppressors regulating gene expression; (4) cREs ± 500bases: extended cREs by 500 bases in both directions. The rationale behind this approach is that different GWAS variants can influence phenotypes by regulating gene expression in a cell-type specific manner through various regulatory mechanisms. For example, they may alter enhancer function (cREs category) or affect the binding of a transcription factor at a gene’s promoter (Promoter OCRs category).

• Download figure
• Open in new tab

Figure 1: Partitioned Linkage Disequilibrium Score Regression analysis for open chromatin regions of all cell types.

A. The schematic shows the different types of regions defined in our study and 3 different ways overlapping chromatin contact regions – OCRs – gene promoters define cREs.

B. Heritability enrichment by LDSC analysis for each cell type

• a) Bar-plot shows the total number of OCRs identified by ATAC-seq for each cell type on bulk cells - blue, or on single cells – red; and the portion of OCRs that fall within cREs identified by incorporating Hi-C (green) or by Capture-C (orange).

• b-e) 4 panels of dot-plots show heritability enrichment by LDSC analysis for each cell type, with standard error whiskers. Dots’ colors correspond to -log10(p-values), dots with white asterisks are significant p-values <0.05, and dots’ sizes corresponding to the proportion of SNP contribute to heritability. Dash line at 1, i.e., no enrichment.

  • b) Analysis done on whole OCRs set of each cell type (whiskers colors match with bulk/single cell from bar-plot a);

  • c) On only OCRs that overlapped with promoters (whiskers’ colors match with bulk/single cell from bar-plot a);

  • d) On the putative cREs of each cell type (whiskers’ colors match with Hi-C/Capture-C from bar-plot a);

  • e) On the same cREs as (c) panel with their genomic positions expanded ±500 bases on both sides (whiskers’ colors match with Hi-C/Capture-C from bar-plot a)

We observed that 41 of 57 cell types – including 22 metabolic, 21 immune, 7 neural cell types and 7 independent cell lines (eTable 1) – showed at least a degree of directional enrichment with the total set of OCRs (Fig. 1B – Total OCRs). However, only four cell types – two pancreatic alpha and two pancreatic beta cell-based datasets – had statistically significant enrichments (P<0.05). These enrichments were less pronounced when focusing on promoter OCRs only (Fig. 1B – Promoter OCRs). To further limit the LD enrichment assessment to just those OCRs that can putatively regulate gene expression via chromatin contacts with gene promoters, we used the putative cREs^27,29. This reduced the number of cell types showing at least nominal enrichment (31 of 57), enlarged the dispersion of enrichment ranges across different cell types, increased the 95% confidence intervals (CI) of enrichments, and hence increased the P-value of the resulting regression score. cREs from pancreatic alpha cells derived from single-cell ATAC-seq were the only dataset that remained statically significant (Fig. 1B – putative cREs).

The original reported LDSR method analyzed enrichment in the 500bp flanking regions of their regulatory categories³⁹. However, when we expanded our analysis to the ±500bp window for our cREs, albeit incorporating more weighted variants into the enrichment (represented by larger dots in Fig. 1B – cREs ±500 bases), this resulted in a decrease in the number of cell types yielding at least nominal enrichment (26 cell types), the enrichment range across cell types, the 95% CI, and level of significance. The pancreatic alpha cell observation also dropped below the bar for significance with this expanded window definition.

Consistency and diversity of childhood obesity proxy variants mapped to cREs

Despite the enrichments above only being limited to just a small number of cell types, it is likely that individual loci have differing levels of contributions in various cellular contexts and could not be detected at the genome wide assessment scale. As such we elected to further explore the candidate effector genes that are directly affected by cREs harboring childhood obesity-associated variants by systematically mapping the genomic positions of the LD proxies onto each cell type’s cREs. Most proxies fall within chromatin contact regions (blue area in Venn diagram Fig. 2A) or OCRs (yellow area) or open chromatin contact regions (red area), or completely outside (white area) any defined region. Only 94 proxies fall within our defined cREs (overlapped area with dotted green border in Fig. 2A), they clustered at 13 original loci (eTable 3). eFig. 1A outlines the number of signals at each locus included or excluded based on the criteria we defined for our regions of interest. The TMEM18 locus yielded the most variants through cREs mapping, with 46 proxies for the two lead independent variants, rs7579427 and rs62104180. The second most abundant locus was ADCY3, with 21 proxies for lead variant rs4077678 (Fig. 2B). The higher number of variants at one locus did not correlate with implicating more genes or cell types through mapping. The mapping frequency of various variants within a specific locus exhibited substantial differences.

• Download figure
• Open in new tab

Figure 2: Mapping 771 proxies to the open chromatin regions of each cell type

A. Venn diagram shows how 771 proxies mapped to the OCRs:

• Blue area: 758 proxies were located within contact regions of at least one cell type regardless of chromatin state;

• Red area: 417 proxies were located within contact regions marked as open by overlapping with OCR;

• Yellow area: If we only considered open chromatin regions, 178 proxies were included;

• Dotted green bordered area: To focus on just those variants residing within open chromatin and contacting promoter regions in any cell type, we overlapped the genomic positions of these proxies with each cell type’s cRE set, yielding 90 variants (3 from the 99% credible set) directly contacting open gene promoters (eTable 3), with 10 of which located within a promoter of one gene but contacting another different gene promoter. There were an additional 4 variants located within gene promoters but in chromatin contact with promoter(s) of nearby transcript(s) of the same gene (correspond to 3 cREs illustrations in Figure 1A).

• White area: proxies that fall into neither defined region of interest. B. Bar-plot shows number of proxies, cell types and target genes mapped at each locus. C. The upSet plot shows the degree of overlap across cell types of the variants; ranked from the most common variant (red) – rs61888800 from BDNF locus, a well-known 5’ untranslated region variant of this gene that is associated with anti-depression and therapeutic response^81,82 – appeared in 39 cell types, to the group of variants (grey) which appeared in only one cell type.

Inspecting individual variants regardless of their locus, we found that 28 of 94 proxies appeared in cREs across multiple cell types, with another 66 observed in just one cell type (Fig. 2D). 45 variants of these 66 just contacted one gene promoter, such as at the GPR1 and TFAP2B loci (eFig. 2).

Overall, the number of cell types in which a variant was observed in open chromatin correlated with the number of genes contacted via chromatin loops (eFig. 3A). However, we also observed that some variants found in cREs in multiple cell types were more selective with respect to their candidate effector genes (eFig. 3B-red arrow), or conversely, more selective across given cell types but implicated multiple genes (eFig. 3B-blue arrow). eFig. 4 outlines our observations at the TMEM18 locus – an example locus involved in both scenarios.

• Download figure
• Open in new tab

Figure 3: Profiles of 111 implicated genes by 94 proxies through cREs of each cell type

Main panel: Bubble plot show corresponding expression level (size) and number of variants (color) target each implicated gene of each cell type. Squares represent genes with variants at their promoters. Circles represent genes with variants contacted through chromatin loops. Some genes were implicated by both types, these “double implications” are represented as diamond shapes, and were identified across several cell types: two cell types (plasmacytoid dendritic cells and pre-differentiated adipocytes) for ADCY3 gene, and five for BDNF (human embryonic stem cells - hESC, differentiated human fetal osteoblast cells - hFOB_Diff, neural progenitor cells derived from induced pluripotent stem cells - NPC_iPSC, PANC-1, and NCIH716 cell lines) Genes with expression undetected in our arrays are shown as triangles.

Top panel: bar-plot shows numbers of cell types each gene was implicated within, color-coded by which systems the cell types belong to.

Right panel: bar-plot shows numbers of genes implicated by the variants with each cell type.

• Download figure
• Open in new tab

Figure 4: Colocalization of target effector genes with eQTLs

A. Venn diagram shows the overlaps between sets of genes yielded by ColocQuiaL and the variant-to-gene mapping process.

B. Circos plot of the 10 loci demonstrates the differences in the ranges of associations between the two approaches, with long-ranged chromatin contacts between obesity variants and target genes displayed as orange links and short-range eQTLs colocalizations as green links.

• Two SNPs – rs35796073, and rs35142762 within the TMEM18 locus, in linkage disequilibrium with rs7579427 – were estimated with high probability (cond.PP.H4=0.78) of colocalizing with the expression of ALKAL2 gene in subcutaneous adipose tissue. These pairs of SNP-gene were also identified by our variant-to-gene mapping approach in natural killer cells, plasmacytoid dendritic cells, unstimulated PBMC naïve CD4 T cells and astrocytes.

• The rs7132908 variant at the FAIM2 locus colocalized with the expression of AQP6 in thyroid tissue and with ASIC1 in prostate tissue, not only with high cond.PP.H4 but also with high individual SNP causal probability (SNP.PP.H4 > 0.95). rs7132908 was the second most consistent observation in our variant-to-gene mapping, namely across 25 different cell types (Figure 2B) and all three systems plus the other independent cell lines. The pair of rs7132908-contacting-AQP6 was observed in 15 different cell types - 8 metabolic and 4 neural cell types, and 3 independent cell lines. The pair of rs7132908-contacting-ASIC1 was observed in 11 different cell types - 8 metabolic and 2 neural cell types, and plasmacytoid dendritic cells.

• The other eQTL signals that overlapped with our variant-to-gene mapping results were: BDNF at the METTL15 locus with its promoter physically contacted by rs11030197 in 4 cell types and its expression significantly colocalized (cond.PP.H4=0.82) in tibial artery; ADCY9 at its locus with its promoter physically contacted by rs2531995 in natural killer cells and its expression significantly colocalized in skin tissue (“Skin_Not_Sun_Exposed_Suprapubic”, cond.PP.H4=0.97). And ADCY3 in the C panel.

C. ColocQuiaL estimated that these SNPs highly colocalize with the expression of ADCY3 in 11 different tissues, where the overlapping with the 16 cell types is represented, color-coded by the proxies rs numbers.

Implicated genes cluster at loci strongly associated with childhood obesity consistently across multiple cell types

Mapping the variants across all the cell types resulted in a total of 111 implicated childhood obesity candidate effector genes (Table 1). Among these, 45 genes were specific to just one cell type (eFig. 5A), including 13 in myotubes and 7 in natural killer cells. Conversely and notably, BDNF appeared across 42 different cell types. Across the metabolic, neural, and immune systems and seven other cell lines, there were 9 genes consistently implicated in all four categories (top panel Fig. 3 – red stars, eFig. 5B: “all”), while 5 genes were consistently implicated in metabolic, neural, and immune systems (top panel Fig. 3 – blue stars, eFig. 5B: “all_main”). Two genes, ADCY3 and BDNF, had variants both at their promoters and contacted variants in cREs via chromatin loops (eFig. 6).

At the TMEM18 locus on chr 2p25.3, a highly significant human obesity locus that has long been associated with both adult and childhood obesity, we obeserved differing degrees of evidence for 16 genes, but noted that rs6548240, rs35796073, and rs35142762 consistently contacted the SH3YL1, ACP1, and ALKAL2 promoters across multiple cell types (Fig. 2D-third and fourth column).

At the chr 2p23 locus, ADCY3 yielded the most contacts (i.e. many proxies contacting the same gene via chromatin loops), suggesting this locus acts as a regulatory hub. However, we observed a similar composition in cell types for four other genes: DNAJC27, DNAJC27-AS1 (both previously implicated in obesity and/or diabetes traits⁴⁰), AC013267.1, and SNORD14 (RF00016). ITSN2, NCOA1, and EFR3B were three genes within this locus that were only implicated in immune cell types. NCOA1 encodes a prominent meta-inflammation factor⁴¹ known to reduce adipogenesis and shift the energy balance between white and brown fat, and its absence known to induce obesity⁴².

CALCR was the most frequently implicated gene at its locus, supported by 20 cell types across all systems. While within the BDNF locus, METTL15 and KIF18A – two non-cell-type-specific genes - plus some lncRNA genes, were contacted by childhood obesity-associated proxies within the same multiple cell types as BNDF, again suggesting the presence of a regulatory hub.

At the FAIM2 locus on chr 12q13.12, we observed known genes associated with obesity, eating patterns, and diabetes-related traits, including ASIC1, AQP2, AQP5, AQP6, RACGAP1, and AC025154.2 (AQP5-AS1) along with FAIM2 (Table 1). These genes were harbored within cREs of astrocytes, neural progenitors, hypothalamic neurons, and multiple metabolic cell types. Plasmacytoid and CD1c+ conventional dendritic cells were the only two immune cell types that harbored such proxies within their cREs, implicating ASIC1, PRPF40B, RPL35AP28, TMBIM6, and LSM6P2 at the FAIM2 locus.

The independent ADCY9 and FTO loci are both located on chromosome 16. Genes at the ADCY9 locus were only implicated in a subset of immune cell types. Interestingly, genes at the FTO locus were only implicated in Hi-C datasets (as opposed to Capture C), including 6 metabolic cell types and astrocytes. Most genes at the FTO locus were implicated in skeletal myotubes, differentiated osteoblasts, and astrocytes, namely FTO and IRX3; while IRX5, CRNDE, and AC106738.1 were also implicated in adipocytes and hepatocytes.

The most implicated cell types by two sets of analyses

EndoC-BH1 and myotubes are the two cell types in which we implicated the most effector genes, with 38 and 42, respectively – Fig. 3 side panel. This phenomenon is likely proportional in the case of myotubes, given the large number of cREs identified by overlapped Hi-C contact data and ATAC-seq open regions (Fig. 1A), but not for EndoC-BH1. Albeit harboring an average number of cREs compared to other cell types, EndoC-BH1 cells were consistently among the top-ranked heritability estimates for the childhood obesity variants resulting from the EGG consortium GWAS (Fig. 1) and harbored a significant number of implicated genes by the mapping of proxies. Interestingly, the pancreatic alpha cell type – shown above to be the most significant for heritability estimate by LDSC – revealed only 6 implicated genes contacted by the defined proxies, namely BDNF and five lncRNA genes.

Pathway analysis

Of the 111 implicated genes in total, PubMed query revealed functional studies for 66 genes. The remaining were principally lncRNA and miRNA genes with currently undefined functions (Table 1). To investigate how our implicated genes could confer obesity risk, we performed several pathway analyses keeping them either separated for each cell type or pooling into the respective metabolic, neural, or immune system sets. eFig. 7 shows simple Gene Ontology (GO) biological process terms enrichment results.

Leveraging the availability of our expression data generated via RNA-seq (available for 46 of 57 cell types), we performed pathway analysis. Given that our gene sets from the variant-to-gene process was stringently mapped, the sparse enrichment from normal direct analyses is not ideal for exploring obesity genetic etiology. Thus, we incorporated two methods from the pathfindR package⁴³ and our customized SPIA (details in eMethods). The result of 60 enriched KEGG terms is shown in eFig. 8 (eTable 4), with 13 genes in 14 cell types for pathfindR and 39 enriched KEGG terms shown in eFig. 9 (eTable 5), with 10 genes in 42 cell types for customized SPIA. There were 20 overlapping pathways between the two approaches (yellow rows in eTable 4&5) including many signaling pathways such as the GnRH (hsa04912), cAMP (hsa04024), HIF-1 (hsa04066), Glucagon (hsa04922), Relaxin (hsa04926), Apelin (hsa04371), and Phospholipase D (hsa04072) signaling pathways. They were all driven by one or more of these 5 genes: ADCY3, ADCY9, CREBBP, MMP2, and NCOA1. Interestingly, we observed the involvement of natural killer cells in nearly all the enriched KEGG terms from pathfindR due to the high expression of the two adenylyl cyclase encoded genes, ADCY3 and ADCY9, along with CREBBP. The SPIA approach disregarded the aquaporin genes (given they appear so frequently in so many pathways that involve cellular channels) but highlighted the central role of BDNF which single-handedly drove four signaling pathways: the Ras, Neurotrophin, PI3K-Akt, and MAPK signaling pathways. This also revealed the role of TRAP1 in neurodegeneration.

These two approaches did not discount the role of FAIM2 and CALCR. However, their absence was mainly due to the content of the current KEGG database. On the other hand, these approaches accentuated the role of the MMP2 gene at the FTO locus in skeletal myotubes, given its consistency within the GnRH signaling pathway (eFig. 10), which is in line with previous studies linking its expression with obesity^44–46.

Supportive evidence by colocalization of target effector genes with eQTLs

The GTEx consortium has characterized thousands of eQTLs, albeit in heterogeneous bulk tissues⁴⁷. To assess how many observed gene-SNP pairs agreed with our physical variant-to-gene mapping approach in our multiple separate cellular settings, we performed colocalization analysis using ColocQuiaL³⁸.

282 genes were reported to be associated with the variants within 13 loci from our variant-to-genes analysis. We found 114 colocalizations for ten of our loci that had high conditioned posterior probabilities (cond.PP.H4.abf ≥ 0.8), involving 44 genes and 41 tissues among the eQTLs. We extracted the posterior probabilities for each SNP within each colocalization and selected the 95% credible set as the likely causal variants (complete list in eTable 6). Despite sensitivity differences and varying cellular settings, when compared with our variant-to-gene mapping results, colocalization analysis yielded consistent identification for 21 pairs of SNP-gene interactions when considering the analyses across all our cell types, composed of 20 SNPs and 7 genes. Details of these SNP-gene pairs are shown in Fig. 4A and B.

Of these 20 SNPs, 15 were at the ADCY3 locus, in LD with sentinel variant rs4077678, and all implicated ADCY3 as the effector gene in 29 cell types – 15 metabolic, 6 immune, 4 neural cell types and 4 independent cell lines (Fig. 4C). Indeed, missense mutations have been previously reported for this gene in the context of obesity^48,49 while another member of this gene family, ADCY5, has also been extensively implicated in metabolic traits⁵⁰.

Predicting transcription factors (TFs) binding disruption at implicated genes contributing to obesity risk

TFs regulate gene expression by binding to DNA motifs at enhancers and silencers, where any disruption by a SNP can potentially cause dysregulation of a target gene. Thus, we used motifbreakR (R package) to predict such possible events at the loci identified by our variant-to-gene mapping. Each variant was predicted to disrupt the binding of several different TFs, thus requiring further literature cross-examination to select the most probable effects. For example, rs7132908 (consistently contacting FAIM2 in 25 cell types) was predicted to disrupt the binding of 12 different transcription factors. Among them, SREBF1 (eFig. 11A) was the only TF that concurred with evidence that it regulates AQP2 and FAIM2 at the same enhancer⁵¹. The full prediction list can be found in eTable 7.

To narrow down the list of putative TF binding sites at each variant position, we leveraged the ATAC-seq footprint analysis using the RGT suite⁵². The final set of Motif-Predicted Binding Sites (MPBS) within each cell type ATAC-seq footprints was used to overlap with the genomic locations of the OCRs, and then overlapped with our obesity variants, resulting in annotated 29 variants. Mosaic plot in eFig. 11B shows the number and proportions of variants predicted by motifbreakR and/or overlapped with MPBS. Insignificant P-value from Fisher’s exact test indicated the independence of the two analyses. Only seven variants were found within the cREs for the same TF motifs predicted to be disrupted by motifbreakR (eFig. 11C). eFig. 12 outlines the seven variants that motifbreakR and ATAC-seq footprint analysis agreed on the TF bindings they might disrupt.