Imagine trying to talk to your boss while your least favorite colleague whispers abuse in your ear.
Envision a repetitive stream of verbal abuse rampaging in your mind — one over which you may have no control, one from an unknown entity, and one which firmly obstructs your day-to-day routine. Suppose that a deep coat of these voices was to mount a challenge to your every thought, or perhaps command your body to realize certain tasks.
This harsh experience captures what the life of an individual with schizophrenia might resemble — one shared by approximately 3.2 million Americans, and the same proportion of people worldwide.
A Window of Opportunity
Defined as a serious mental disorder (with a heterogeneous presentation and variable outcome) in which people interpret reality abnormally, the active symptoms of schizophrenia (SZ) can comprise marked impairments in multiple domains necessary for proper daily functioning, affecting the ability of a substantial proportion of patients to maintain social relationships, sustain employment, and live independently.
The following graphic thoroughly outlines the onset and progression of schizophrenia in relation to risk factors. The initial diagnosis of schizophrenia, which operationally corresponds to the patient’s first episode of psychosis (FEP), is typically made in young adulthood, generally following a prodromal, at-risk phase in which sub-threshold psychotic episodes and other symptoms are noticed. Although approximately 10–15% of patients recover after their FEP, a similar proportion exhibit a more unremitting form of the disorder — one described to follow “a fluctuating course punctuated by acute exacerbation of psychotic crises superimposed upon a background of poorly controlled negative, neurocognitive and social cognitive symptoms.”
The graph below covers the developmental processes affected by the disorder, the course to and progression of which can be related to 3 fundamental phases in the “life of the brain”: brain formation, reorganization, and upkeep. The latter two of these phases in particular encompasses a range of processes which could be suitable for therapeutic intervention. In fact, in the absence of an outright cure for the disorder, the most compelling “window of opportunity” to impede onset or block early progression in the eyes of experienced psychiatrists is around the FEP, as seen here:
The Chromosomal Connectome
Connectomics — a big data approach for the analysis of the massive datasets produced by functional and structural brain imaging — offers a valuable lens through which the disease’s effects can be understood. The following graphic portrays that patients with schizophrenia showed decreased functional connectivity in two distinct brain networks, alongside a significant breakdown in the normal relationship between functional and structural brain connectivity:
Of course, whilst the connectomic disturbances in patients with schizophrenia have been comprehensively mapped across the entire brain, its developmental reorganization in the aforementioned “window of opportunity” remains beyond the scope of conventional connectomic exploration.
Consider spatial genome organization — highly regulated and foundational to normal brain development and function. Indeed, many of the risk variants contributing to the heritability of complex genetic psychiatric disorders (including schizophrenia) are likely embedded in “three-dimensional genome” (3DG) structures in non-coding sequences important for transcriptional regulation. For instance, these structures include chromosomal loop formations that bypass linear genome on a kilobase/megabase scale and topologically associated domains (TADs).
Introducing Neurogram
By linking schizophrenia-associated genetic variants with distal gene targets, 3DG mapping with Hi-C (and other) genome-scale approaches could inform how higher-order chromatin organization affects the genetic risk for schizophrenia, paving the path for neuroscientists to identify vulnerable patients and psychiatrists to impede early progression of the psychiatric disease.
Indeed, in the pursuit of exploring the developmental reorganization of the three-dimensional genome of the brain in the context of neuropsychiatric disease, the chromosomal conformations in differentiating neural progenitor cells can be monitored. The techniques with which this project is executed are dubbed Neurogram (a packaging label of sorts, for simplicity).
Special Thanks
This project is a replication of the following paper, the authors of which I owe a debt of gratitude for providing the code and data used, as well as informing the corresponding theory.
Prashanth Rajarajan, Tyler Borrman, Will Liao, et al. (2018). Neuron-specific signatures in the chromosomal connectome associated with schizophrenia risk. Science.
Rationale
It ought to be noted that 3DG mapping from post-mortem tissue is devoid of cell type–specific resolution, and so it might not capture higher-order chromatin structures sensitive to the autolytic process. Accordingly, the developmentally regulated changes in chromosomal conformations during the course of isogenic neuronal and glial differentiation are to be monitored, allowing us to delineate the large-scale pruning of chromosomal contacts in the transition from neural progenitor cells (NPCs) to neurons.
Moreover, these methods enable us to unearth an expanded 3DG risk space for schizophrenia — one with a functional network of disease-specific regulators of neuronal connectivity, synaptic signaling, and chromatin remodeling.
Linking Neural Differentiation to Large-scale 3DG Remodeling
To begin with, researchers applied in situ Hi-C to map the three-dimensional genome of two male human induced pluripotent stem cell (hiPSC)–derived neural progenitor cells. This genome-scale approach is further applied to isogenic populations of induced excitatory neurons (which were generated through viral overexpression of the Neurogenin 2 (NGN2) transcription factor) as well as differentiations of astrocyte-like glial cells.
The upper portion of the figure below depicts the derivation scheme of isogenic cell types from two male control cell lines, where the pink ovals correspond to donor hiPSCs, the orange cells represent NPCs, and the green and purple cells depict neurons and glia respectively:
The lower section, on the other hand, constitutes the hierarchical clustering of intra-chromosomal interactions from six in situ Hi-C libraries, where height signifies the distance between libraries.
The second figure below illustrates the immunofluorescent staining of characteristic cell markers for NPCs (Nestin and SOX2), neurons (TUJ1 and MAP2), and glia (Vimentin and S100β):
Transcriptome RNA sequencing comparison with published datasets subsequently validates that the NPCs, but not glia, from subjects S1 and S2 clustered together with NPCs from independent donors, while S1 and S2 NGN2 neurons were found to be closely aligned with directed differentiation forebrain neurons and prenatal brain datasets. Both the transcriptomic datasets and the hierarchical clustering of post-processing Hi-C datasets showed distinct separation by cell type.
Genome-scale interaction matrices were then enriched for intra-chromosomal conformations (excluding the negative control NPC library, in which the ligase step was omitted). On account of the observed correlation between technical replicates of Hi-C assays from the same cell type and donor, alongside the correlation between cell type–specific Hi-C from the two donors, pooling by cell type was conducted.
Conservatively defined as “distinct contacts between two loci in the absence of similar interactions in the surrounding sequences”, intra-chromosomal loop formations were targeted first. Subsequent comparative analyses — including published in situ Hi-C data from the B lymphocyte–derived cell line GM12878 — with the HiCCUPS pipeline (with 5- and 10-kb loop resolutions in conjunction, subsampled to 372 million valid intra-chromosomal read pairs) called 17,767 distinct loops.
The figure above depicts a Venn diagram of these loop calls — one in which 5,068 (28.5%) were specific to only one of the four cell types and 3,118 (17.5%) were shared among all four cell types.
The following table covers the gene (GO) enrichment of genes overlapping anchors of loops shared by NPCs, (brain-specific) neurons, and glia, but absent in the B lymphocyte–derived cell line GM12878:
The inclusion of terms unidentified in lymphocytes, including “forebrain development” and “neuron differentiation”, attests to strong tissue-specific loop signatures.
Surprisingly, a reduction in the chromosomal loop count in neurons relative to isogenic glia and NPCs was identified, alongside a decreased density of chromosomal conformations in genome browser visualization of chromosomal arms, such as chr17q.
The figure above captures cell-type pooled whole-genome heatmaps at 500-kb resolution.
The following figure, by contrast, portrays arc maps showing intra-chromosomal interactions at 40-kb resolution of the q-arm of chr17 for isogenic neurons, NPCs, and glia, where green, orange, and purple colors correspond to neurons, NPCs, and glia respectively. RNA sequencing tracks are shown above the arc maps for each cell type:
While approximately 13,000 loop formations were found in both NPCs and glia, neurons harbored just 7,206, 442 of which were neuron-specific loop formations. One such example includes CUX2 — a transcription factor highly expressed in the NGN2-induced neurons.
Indeed, the figure below presents the FPKM gene expression of CUX2 across the three different cell types, each with heatmaps in which the black arrow indicates the CUX2 loop formation:
It was observed that NPCs, neurons, and glia had similar proportions of loops anchored in solely active compartments (indicated by “A”), solely inactive compartments (indicated by “B”), or in both, signaling no preferential loss of either sort of loop in neurons.
This is illustrated in the following figure, which depicts the number of loops specific to each cell type with one anchor in an A compartment and another in a B department, both in either B compartments (shown in red) or A compartments (shown in blue):
However, from the group of genes overlapping anchors of loops that underwent pruning during the NPC-to-neuron transition, regulators of cell proliferation, morphogenesis, and neurogenesis ranked prominently in the top 25 GO terms with significant enrichment — a finding consistent with a departure from the precursor stage toward post-mitotic neuronal identity.
Aiming to confirm that the observed net loss of loop formations during this transition could be replicated across an array of in vivo approaches and cell cultures, additional Hi-C experiments were conducted on cells differentiated from hiPSC-NPCs via non-NGN2 protocols. A further group of Hi-C datasets generated from a mouse model of neural differentiation were analyzed, consisting of mouse embryonic stem cells (mESCs), mESC-derived NPCs (mNPC), and cortical neurons (mCN) differentiated from the mNPCs via inhibition of the Sonic Hedgehog (SHH) pathway. Moreover, Hi-C data from the human fetal cortical plate (CP) and forebrain germinal zone (GZ) were reanalyzed, to determine if genome-wide chromosomal loop remodeling occurred in the developing brain in vivo.
A 20% decrease in loops compared with their neural progenitors was observed across both the hiPSC-NPC-to-forebrain neuron and mESC-mNPC-mCN differentiation, while a consistent supplementary finding derived that in vivo CP compared with GZ exhibited a 13% decrease in loops genome-wide. Beyond simply harboring fewer total loops, though, neurons displayed a higher proportion of longer-range loops (defined as > 100 kb) than NPCs or glia did. These results establish that in humans and mice, neural precursors to young neurons reliably evidenced a reduced number of loops in neuron-enriched tissues and cultures.
Although overall TAD landscapes remained similar between neurons, NPCs, and glia — a result consistent with relevant studies — TADs also showed an increase of approximately 10% in average neuron size compared with isogenic NPCs, irrespective of which differentiation protocol was applied, as established in a Wilcoxon-Mann-Whitney test:
The figure above provides a box-and-whisker distribution plot of TAD size across four cell types (left), with median TAD length for each type also shown (right).
The aforementioned 10% increase is highlighted in the following figure’s heatmaps at 40-kb resolution for a 3-Mb window at the CDH2 locus on chr18. The figure encompasses a nested TAD landscape in glia with multiple subTADs (indicated by black arrows), which is absent from neuronal Hi-C. The prior color-coding scheme in which green and purple cells signify neurons and glia respectively is maintained:
Based on these findings, TAD remodeling may correspond to restructuring of nested subdomains within larger neuronal TADs. In examining whether this form of developmental reorganization of the brain’s spatial genomes was associated with a generalized shift in chromatin structure, the assay for transposase accessible chromatin with high-throughput sequencing was applied to map open chromatin sequences before and after NGN2-neuronal induction.
The minimal difference between NPCs and neurons in the genome-wide distribution profiles for transposase-accessible chromatin, alongside the reasonably low chromatin accessibility of –2.5 for ≥89% of the anchor sequences comprising cell type–specific and shared brain loops in the cell culture system, speak to widespread 3DG changes during the NPC-to-neuron transition and NPC-to-glia transition in the brains of humans and mice — changes unlikely to be accounted for by global chromatin accessibility differences.
Mapping Chromosomal Contacts at Schizophrenia Risk Loci
Since numerous schizophrenia risk variants are found in non-coding regions in proximity to several genes, chromosomal contact mapping bears the potential to resolve putative regulatory elements with the ability to confer schizophrenia risk via their physical proximity to the target gene. As such, our cell type–specific interactions can be overlaid onto the 145 risk loci associated with schizophrenia risk. Interactions are defined as filtered contacts noticeable over the global background and binomial statistics is applied to identify chromosomal contacts anchored at disease-relevant loci; this approach ensures a maximally comprehensive exploration of the 3DG in context of schizophrenia-related sequences.
First, the 40 loci with the strongest statistical evidence for co-localization of an adult post-mortem brain eQTL and schizophrenia genome-wide association study (GWAS) signal are examined. It is determined that approximately 30% of these risk locus–associated eQTLs bypass the linear genome and are located near the proximal promoter and transcription start site of the target gene.
When queried for schizophrenia-associated risk loci, cell type–specific contact maps with 10-kb-wide bins often unveiled differential chromosomal conformations in NPCs, glia, and neurons. One such example involves the risk locus upstream of the PROTOCADHERIN (PCDH) cell adhesion molecule gene clusters (chromosome 5) — one with great relevance to neuronal connectivity in both the developing and adult brain.
It was found to show a bifurcated bundle of interactions in NPCs through both observed/expected interaction matrices and global background-filtered contact mapping; while one bundle emanated to sequences 5′, the other bundle advanced to sequences 3′ from the locus. In neurons, on the other hand, the 3′ bundle was maintained and the 5′ bundle was “pruned,” whereas glia exhibited the opposite pattern.
The figure above portrays observed/expected interaction heatmaps at 10-kb resolution for the risk-associated clustered PCDH locus for NPC, glia, and neurons.
The following grayscale heatmap, by contrast, depicts areas of highly cell-specific contact enrichments; this entails upstream genes and downstream PCDH gene clusters (denoted by “A” and “B” respectively):
The figure below depicts violin plots of the aforementioned observed/expected interaction values in the regions A and B, highlighting the highly significant differences between the three cell types:
Finally, the following figure renders the map of chromosomal contacts identified by binomial statistics, where the red box with the dashed black lines represents the schizophrenia risk locus:
The dosage of the non-coding schizophrenia risk-SNP (single nucleotide polymorphism) greatly increased the expression of genes that were interconnected to the disease-relevant non-coding sequence in neurons and NPCs (but not in glia). Cell type–specific Hi-C thus identified chromosomal contacts anchored in schizophrenia-associated risk sequences with the ability to affect expression of the target genes.
The transcriptional profiles hiPSC-derived neurons and NPCs closely resembled those of a human fetus during the first trimester, while components of the genetic risk architecture of schizophrenia matched regulatory elements observed to be highly active during pre-natal development. Accordingly, in the Hi-C datasets, seven loci were examined, which contained 36 potentially causal schizophrenia-risk SNPs; all such loci harbored chromosomal interactions in the fetal brain with genes underlying neuron development and function.
Observing the conservation of risk-associated chromosomal contacts between the hiPSC-NPCs, human fetal CP, and germinal zone Hi-C datasets for 5 of the 7 loci observed, in order to test the regulatory function of these conserved risk sequence-bound conformations, single-guide RNA–based epigenomic editing experiments were performed on certain isogenic antibiotic-selected NPCs. ASCL1-, EFNB1-, MATR-3, and SOX2- bound chromosomal contacts separated by 200- to 700-kb interspersed sequences were tested, the results of which are presented below:
The figure above shows cell-type resolved contact maps of 10-kb bins (denoted by the black vertical lines) within risk sequences on chr12 (left), chrX (middle), and chr5 (right), while gene models are outlined (bottom) with the SNP-loop target gene highlighted (red). Again, the green, orange, and purple colors correspond to neurons, NPCs, and glia respectively, whereas the –log q value captures the significance of the contact between the schizophrenia risk locus and each 10-kb bin.
The following figure covers the epigenomic editing for three risk SNP-target gene pairs and their respective control sequences (top), measured with reverse transcription polymerase chain reactions (RT-PCR) for the VP64 (middle) and VPR (bottom) transcriptional activators:
The epigenomic editing of risk sequences 500 to 600 kb distant from the SOX2 and MATR3 loci was found not to alter target gene expression. Since elements of the MATR3-bound risk sequences are embedded in repressive chromatin, five sgRNAs for Cas9 nuclease mutagenesis were directed toward a 138–base pair (bp) sequence, aiming to disrupt it. This technique produced a substantial increase in MATR3 expression upon ablation of the putative repressor sequence.
After conducting additional genomic mutagenesis assays, Cas9 nuclease deletion of interacting credible SNPs was found to dramatically increase gene expression of ASCL1, EFNB1, and EP300.
The figure above captures the changes in quantitative PCR gene expression upon directing catalytically active Cas9 to schizophrenia risk-associated credible SNPs (denoted by the vertical red dashes with rsIDs).
A similar targeting approach concerning four credible SNPs upstream of the clustered PCDH locus significantly decreased levels PCDHA8 and PCDHA10 by approximately 50 to 60% — two of the genes whose expression increased with dosage of the risk SNP rs111896713 in the adult post-mortem brain.
These editing assays collectively validate that chromosomal contacts anchored in schizophrenia risk loci can potentially affect target gene expression across hundreds of kilobases; this finding is consistent with developing and adult human brain tissue alongside predictions from chromosomal conformation maps from hiPSC-derived brain cells.
Associating the GWAS Risk Connectome with Gene Co-regulation
All 145 GWAS-defined schizophrenia risk loci can now be investigated together. For clarity, the resulting network of risk loci and their 3D proximal genes are dubbed the “schizophrenia-related chromosomal connectome.”
In this study, neurons and NPCs — but not isogenic glia — displayed a high prevalence of chromosomal contacts with schizophrenia-associated risk loci — there were 1203, 1100, and 425 contacts involving schizophrenia risk sequences that were highly specific to neurons, NPCs, and glia respectively, with mean distances between risk and target bits of 510 kb, 520 kb, and 580 kb respectively.
The figure above shows counts of highly cell type–specific contacts associated with schizophrenia risk in each of the three hiPSC-derived cell types.
This investigation also identified unexpectedly robust cell type– and gene ontology– specific signatures; notably, this group included genes associated with neuronal connectivity and synaptic signaling
The figure above captures the GO enrichment of genes located in schizophrenia risk contacts in NPCs (left), neurons (middle), and glia (right).
Since spatial 3DG proximity of genes serves as an indicator for potential co-regulation, subsequent experiments tested whether the neural cell type–specific schizophrenia-related chromosomal connectome exhibited evidence of coordinated transcriptional regulation and proteomic interaction of the participating genes. To do so, lists of genes anchored in the most highly cell type–specific schizophrenia risk–associated contacts were generated.
The following graphic summarizes the total count of genes (top) for the NPC-, neuron-, and glia-specific contacts, with the number located within the risk loci highlighted (bottom):
Dubbing the intra-chromosomal contact genes found outside of schizophrenia risk loci as “risk locus-connect” genes — specifically those only identified through Hi-C interaction data. Depending on cell type, the current network of known genes overlapping risk sequences (informed only by GWAS) was expended by anywhere from 50 to 150%.
The following 3-part diagram depicts a schematic workflow of analyses performed with cell type–specific contact genes (left), a Venn diagram of genes located in the 145 loci and those found in cell type–specific contacts (middle), and another schematic workflow of analyses, performed with a combined set of “risk locus” and “risk locus–connect” genes (right):
A merged transcriptome dataset was subsequently analyzed in order to assess whether disease-associated cell type–specific chromosomal connectomes were linked to a coordinated program of gene expression. By examining the pair-wise correlations of the collective sets of 386 NPC, 385 neuron, and 201 glia genes representing “risk locus” and “risk locus–connect” genes, it was found that the risk connectome for each cell type showed remarkably strong pair-wise correlations.
Moreover, the averaged gene-by-gene transcript correlation index for each matrix overall (dubbed the “organization score”) for the NPCs, neurons, and glia were 0.22 to 0.25 — representing degrees of organized gene expression seen as robustly significant for NPCs and neurons after controlling for linear genomic distance.
The figure above depicts RNA Pearson transcriptomic correlation heatmaps consisting of risk locus and risk locus–connect genes derived from the cell type–specific contacts of NPCs (left), neurons (middle), and glia (right), with the organization scores (denoted by “|r|avg”) reported for each heatmap along with P values from sampling analysis. The schematics above the heatmaps constitute representations of each cell type’s particular connectome alongside frequency distributions of organization scores from permutation analyses of randomly generated heatmaps, with the red bar indicating the observed organization score of the particular heatmap being tested. Below the heatmap, the gray bar corresponds to n genes with at least 1 count per million in the RNA sequencing dataset out of the total number of genes and use in the construction of the heatmap; the red and blue bars and the corresponding counts below them, on the other hand, designate the number of risk locus (red) and risk locus–connect (blue) genes in the heatmap. Finally, while the fuchsia bars represent synaptic function genes, the yellow bars correspond to chromatin remodeling genes.
Based on these findings, one can conclude that the chromosomal connectomes associated with schizophrenia risk are indeed cell type–specific, with the neuronal risk connectome appearing particularly enriched for genes related to neuronal connectivity, synaptic signaling, and chromatin remodeling.
Associating the GWAS Risk Connectome with Protein-protein Association Networks
Many proteins encoded by risk locus and risk locus–connect genes were found to be associated with synaptic signaling — these two classes of genes exhibited major protein-protein interaction network effects for NPCs and neurons, but not for glia.
The figure above depicts an overview of protein-protein association networks in NPCs (left), neurons (middle), and glia (right), along with the number of edges connecting the proteins in each network and computed P values. While the gray bars indicate the subset of genes whose proteins are involved in the network out of the total number of genes from cell type–specific interactions, the indication procured by the red and blue bars (and their corresponding counts) remain unchanged — namely, how how many of the genes in the network are in a risk locus (red) and are risk locus–connect (blue).
When compared with randomly generated subset heatmaps from the full schizophrenia-related chromosomal connectome, the transcriptomic correlation heatmaps for these protein networks (dubbed “STRING” genes) had higher organization scores in NPCs and neurons.
The graph above represents a comparison of the organization scores between the full RNA transcriptomic correlation heatmaps (brown) and the “STRING” heatmaps (tan).
Since the removal of the NPC STRING protein network genes significantly decreased the transcriptomic correlation heatmap for the schizophrenia-related chromosomal connectome, this subset of STRING-interacting proteins may drive the observed orchestrated co-regulation. Many instances of co-regulated (RNA) and interacting (protein) risk locus and risk locus–connect genes that share the same TAD were located within these transcriptome- and proteome-based regulatory networks — this includes CDC20, a gene which regulates dendrite development and is associated at the protein level with RNF220, an E3 ubiquitin-ligase and b-catenin stabilizer.
The figure above shows a representative neuronal TAD landscape depicting a schizophrenia risk–associated locus (red) with its risk locus–connect genes (blue) and several genes which are members of the neuronal schizophrenia protein network, which itself is denoted by a fully green circle; the green circle with a gray border, on the other hand, depicts the protein-level interaction between CDC20 and RNF220.
Finally, in order to assess whether such co-regulation could parallel that of the adult human brain’s prefrontal cortex proteome, a newly generated mass spectrometry–based dataset of 182 neuronal proteins was screened. Four of these were from the risk-associated neuronal protein network: GABBR1, GRIA1, GRIN2A, and GRM3. Accordingly, the protein-protein correlation scores were found to be significantly higher for these four risk-associated proteins than expected by random permutation analysis.
The leftmost graphic in the figure above covers the liquid chromatography–selected reaction monitoring (LC-SRM) mass spectrometry (MS) performed on dorsolateral prefrontal cortex (DLPFC) tissue 43 adult post-mortem brains (23 of which belong to patients with schizophrenia, with the remaining 20 being controls). The central illustration summarizes the reliable quantification of 182 neuronal proteins, while the rightmost graph summarizes the findings regarding that the aforementioned four proteins (GABBR1, GRM3, GRIN2A, and GRIA1).
Based on these findings, it can be concluded that the schizophrenia-related chromosomal connectome tethers other portions of the genome to sequences associated with schizophrenia heritability and provides a structural foundation for a functional connectome that reflects coordinated regulation of gene expression and interaction within the proteome.
Conclusions
These experiments validate that neural progenitor differentiation into neurons and glia is associated with the dynamic remodeling of chromosomal conformations and that differentiation-induced loop pruning primarily affects a subset of genes critical for neurogenesis (furnished via NPC-to-neuron loss) and neuronal function (furnished via NPC-to-glia loss).
These analyses also suggest that developmental 3DG remodeling affects a large number of sequences that confer liability for schizophrenia, and reveal the remarkably strong correlation exhibited by genes in 3D physical proximity with schizophrenia-risk variants at the level of the transcriptome and proteome.
Interestingly, the three major functional categories linked to the genetic risk architecture of schizophrenia — neuronal connectivity, synaptic signaling, and chromatin remodeling — had heavy representation within the cell type–specific chromosomal connectomes of neurons and NPCs, and in whole tissue in vivo. As such, cell type–specific 3DG reorganization during the course of neural progenitor differentiation could profoundly impact our understanding of the genetic underpinning of the psychiatric disease.
Closing Thoughts
Millions of people around the world battle through life with voices screeching down their ears. In the absence of an outright cure for the psychiatric disease, early detection and treatment with medicines and psychosocial therapy is considered the most effective form of support.
By adopting an approach within the domain of psychiatric genomics, schizophrenia-associated genetic variants can be linked with distal gene targets. Having monitored the chromosomal conformations in differentiating neural progenitor cells, the disproportionate developmentally regulated changes in neurons highlight the presence of cell type–specific disease risk vulnerabilities in spatial genome organization.
With the effect of higher-order chromatin organization on genetic risk for schizophrenia having been elucidated by 3DG mapping with Hi-C genome-scale approaches, Neurogram thus facilitates the tracking of these risk vulnerabilities in the pursuit of identifying potentially susceptible patients and impeding disease progression before it’s too late.
Special thanks once again to Prashanth Rajarajan, Tyler Borrman, Will Liao, Nadine Schrode, Erin Flaherty, Charlize Casiño, Samuel Powell, Chittampalli Yashaswini, Elizabeth A. LaMarca, Bibi Kassim, Behnam Javidfar, Sergio Espeso-Gil, Aiqun Li, Hyejung Won, Daniel H. Geschwind, Seok-Man Ho, Matthew MacDonald, Gabriel E. Hoffman, Panos Roussos, Bin Zhang, Chang-Gyu Hahn, Zhiping Weng, Kristen J. Brennand, and Schahram Akbarian — the authors of the paper Neuron-specific signatures in the chromosomal connectome associated with schizophrenia risk.
