Identifying patterns differing between high-dimensional datasets with generalized contrastive PCA.

High-dimensional data have become ubiquitous in the biological sciences, and it is often desirable to compare two datasets collected under different experimental conditions to extract low-dimensional patterns enriched in one condition. However, traditional dimensionality reduction techniques cannot accomplish this because they operate on only one dataset. Contrastive principal component analysis (cPCA) has been proposed to address this problem, but it has seen little adoption because it requires tuning a hyperparameter resulting in multiple solutions, with no way of knowing which is correct. Moreover, cPCA uses foreground and background conditions that are treated differently, making it ill-suited to compare two experimental conditions symmetrically. Here we describe the development of generalized contrastive PCA (gcPCA), a flexible hyperparameter-free approach that solves these problems. We first provide analyses explaining why cPCA requires a hyperparameter and how gcPCA avoids this requirement. We then describe an open-source gcPCA toolbox containing Python and MATLAB implementations of several variants of gcPCA tailored for different scenarios. Finally, we demonstrate the utility of gcPCA in analyzing diverse high-dimensional biological data, revealing unsupervised detection of hippocampal replay in neurophysiological recordings and heterogeneity of type II diabetes in single-cell RNA sequencing data. As a fast, robust, and easy-to-use comparison method, gcPCA provides a valuable resource facilitating the analysis of diverse high-dimensional datasets to gain new insights into complex biological phenomena.

Distinct biological signature and modifiable risk factors underlie the comorbidity between major depressive disorder and cardiovascular disease.

Major depressive disorder (MDD) and cardiovascular disease (CVD) are often comorbid, resulting in excess morbidity and mortality. Here we show that CVDs share most of their genetic risk factors with MDD. Multivariate genome-wide association analysis of shared genetic liability between MDD and atherosclerotic CVD revealed seven loci and distinct patterns of tissue and brain cell-type enrichments, suggesting the involvement of the thalamus. Part of the genetic overlap was explained by shared inflammatory, metabolic and psychosocial or lifestyle risk factors. Our data indicated causal effects of genetic liability to MDD on CVD risk, but not from most CVDs to MDD, and showed that the causal effects were partly explained by metabolic and psychosocial or lifestyle factors. The distinct signature of MDD-atherosclerotic CVD comorbidity suggests an immunometabolic subtype of MDD that is more strongly associated with CVD than overall MDD. In summary, we identified biological mechanisms underlying MDD-CVD comorbidity and modifiable risk factors for prevention of CVD in individuals with MDD.

Schizophrenia genomics: genetic complexity and functional insights.

Determining the causes of schizophrenia has been a notoriously intractable problem, resistant to a multitude of investigative approaches over centuries. In recent decades, genomic studies have delivered hundreds of robust findings that implicate nearly 300 common genetic variants (via genome-wide association studies) and more than 20 rare variants (via whole-exome sequencing and copy number variant studies) as risk factors for schizophrenia. In parallel, functional genomic and neurobiological studies have provided exceptionally detailed information about the cellular composition of the brain and its interconnections in neurotypical individuals and, increasingly, in those with schizophrenia. Taken together, these results suggest unexpected complexity in the mechanisms that drive schizophrenia, pointing to the involvement of ensembles of genes (polygenicity) rather than single-gene causation. In this Review, we describe what we now know about the genetics of schizophrenia and consider the neurobiological implications of this information.

Distinct biological signature and modifiable risk factors underlie the comorbidity between major depressive disorder and cardiovascular disease.

Major depressive disorder (MDD) and cardiovascular disease (CVD) are often comorbid, resulting in excess morbidity and mortality. Here we show that CVDs share most of their genetic risk factors with MDD. Multivariate genome-wide association analysis of shared genetic liability between MDD and atherosclerotic CVD revealed seven loci and distinct patterns of tissue and brain cell-type enrichments, suggesting the involvement of the thalamus. Part of the genetic overlap was explained by shared inflammatory, metabolic and psychosocial or lifestyle risk factors. Our data indicated causal effects of genetic liability to MDD on CVD risk, but not from most CVDs to MDD, and showed that the causal effects were partly explained by metabolic and psychosocial or lifestyle factors. The distinct signature of MDD-atherosclerotic CVD comorbidity suggests an immunometabolic subtype of MDD that is more strongly associated with CVD than overall MDD. In summary, we identified biological mechanisms underlying MDD-CVD comorbidity and modifiable risk factors for prevention of CVD in individuals with MDD.

Distinct genomic signatures and modifiable risk factors underly the comorbidity between major depressive disorder and cardiovascular disease.

Major depressive disorder (MDD) and cardiovascular disease (CVD) are often comorbid, resulting in excess morbidity and mortality. Using genomic data, this study elucidates biological mechanisms, key risk factors, and causal pathways underlying their comorbidity. We show that CVDs share a large proportion of their genetic risk factors with MDD. Multivariate genome-wide association analysis of the shared genetic liability between MDD and atherosclerotic CVD (ASCVD) revealed seven novel loci and distinct patterns of tissue and brain cell-type enrichments, suggesting a role for the thalamus. Part of the genetic overlap was explained by shared inflammatory, metabolic, and psychosocial/lifestyle risk factors. Finally, we found support for causal effects of genetic liability to MDD on CVD risk, but not from most CVDs to MDD, and demonstrated that the causal effects were partly explained by metabolic and psychosocial/lifestyle factors. The distinct signature of MDD-ASCVD comorbidity aligns with the idea of an immunometabolic sub-type of MDD more strongly associated with CVD than overall MDD. In summary, we identify plausible biological mechanisms underlying MDD-CVD comorbidity, as well as key modifiable risk factors for prevention of CVD in individuals with MDD.

Transcriptional maintenance of cortical somatostatin interneuron subtype identity during migration.

Although cardinal cortical interneuron identity is established upon cell-cycle exit, it remains unclear whether specific interneuron subtypes are pre-established, and if so, how their identity is maintained prior to circuit integration. We conditionally removed Sox6 (Sox6-cKO) in migrating somatostatin (Sst+) interneurons and assessed the effects on their mature identity. In adolescent mice, five of eight molecular Sst+ subtypes were nearly absent in the Sox6-cKO cortex without a reduction in cell number. Sox6-cKO cells displayed electrophysiological maturity and expressed genes enriched within the broad class of Sst+ interneurons. Furthermore, we could infer subtype identity prior to cortical integration (embryonic day 18.5), suggesting that the loss in subtype was due to disrupted subtype maintenance. Conversely, Sox6 removal at postnatal day 7 did not disrupt marker expression in the mature cortex. Therefore, Sox6 is necessary during migration for maintenance of Sst+ subtype identity, indicating that subtype maintenance requires active transcriptional programs.

Transcriptional diversity in specific synaptic gene sets discriminates cortical neuronal identity.

Synapse diversity has been described from different perspectives, ranging from the specific neurotransmitters released, to their diverse biophysical properties and proteome profiles. However, synapse diversity at the transcriptional level has not been systematically identified across all synapse populations in the brain. To quantify and identify specific synaptic features of neuronal cell types we combined the SynGO (Synaptic Gene Ontology) database with single-cell RNA sequencing data of the mouse neocortex. We show that cell types can be discriminated by synaptic genes alone with the same power as all genes. The cell type discriminatory power is not equally distributed across synaptic genes as we could identify functional categories and synaptic compartments with greater cell type specific expression. Synaptic genes, and specific SynGO categories, belonged to three different types of gene modules: gradient expression over all cell types, gradient expression in selected cell types and cell class- or type-specific profiles. This data provides a deeper understanding of synapse diversity in the neocortex and identifies potential markers to selectively identify synapses from specific neuronal populations.

Single-cell transcriptomics reveals correct developmental dynamics and high-quality midbrain cell types by improved hESC differentiation.

Stem cell technologies provide new opportunities for modeling cells in health and disease and for regenerative medicine. In both cases, developmental knowledge and defining the molecular properties and quality of the cell types is essential. In this study, we identify developmental factors important for the differentiation of human embryonic stem cells (hESCs) into functional midbrain dopaminergic (mDA) neurons. We found that laminin-511, and dual canonical and non-canonical WNT activation followed by GSK3ß inhibition plus FGF8b, improved midbrain patterning. In addition, neurogenesis and differentiation were enhanced by activation of liver X receptors and inhibition of fibroblast growth factor signaling. Moreover, single-cell RNA-sequencing analysis revealed a developmental dynamics similar to that of the endogenous human ventral midbrain and the emergence of high-quality molecularly defined midbrain cell types, including mDA neurons. Our study identifies novel factors important for human midbrain development and opens the door for a future application of molecularly defined hESC-derived cell types in Parkinson disease.

Genome-wide association meta-analysis identifies 48 risk variants and highlights the role of the stria vascularis in hearing loss.

Hearing loss is one of the top contributors to years lived with disability and is a risk factor for dementia. Molecular evidence on the cellular origins of hearing loss in humans is growing. Here, we performed a genome-wide association meta-analysis of clinically diagnosed and self-reported hearing impairment on 723,266 individuals and identified 48 significant loci, 10 of which are novel. A large proportion of associations comprised missense variants, half of which lie within known familial hearing loss loci. We used single-cell RNA-sequencing data from mouse cochlea and brain and mapped common-variant genomic results to spindle, root, and basal cells from the stria vascularis, a structure in the cochlea necessary for normal hearing. Our findings indicate the importance of the stria vascularis in the mechanism of hearing impairment, providing future paths for developing targets for therapeutic intervention in hearing loss.

Postnatal Sox6 Regulates Synaptic Function of Cortical Parvalbumin-Expressing Neurons.

Cortical parvalbumin-expressing (Pvalb+) neurons provide robust inhibition to neighboring pyramidal neurons, crucial for the proper functioning of cortical networks. This class of inhibitory neurons undergoes extensive synaptic formation and maturation during the first weeks after birth and continue to dynamically maintain their synaptic output throughout adulthood. While several transcription factors, such as Nkx2-1, Lhx6, and Sox6, are known to be necessary for the differentiation of progenitors into Pvalb+ neurons, which transcriptional programs underlie the postnatal maturation and maintenance of Pvalb+ neurons' innervation and synaptic function remains largely unknown. Because Sox6 is continuously expressed in Pvalb+ neurons until adulthood, we used conditional knock-out strategies to investigate its putative role in the postnatal maturation and synaptic function of cortical Pvalb+ neurons in mice of both sexes. We found that early postnatal loss of Sox6 in Pvalb+ neurons leads to failure of synaptic bouton growth, whereas later removal in mature Pvalb+ neurons in the adult causes shrinkage of already established synaptic boutons. Paired recordings between Pvalb+ neurons and pyramidal neurons revealed reduced release probability and increased failure rate of Pvalb+ neurons' synaptic output. Furthermore, Pvalb+ neurons lacking Sox6 display reduced expression of full-length tropomyosin-receptor kinase B (TrkB), a key modulator of GABAergic transmission. Once re-expressed in neurons lacking Sox6, TrkB was sufficient to rescue the morphologic synaptic phenotype. Finally, we showed that Sox6 mRNA levels were increased by motor training. Our data thus suggest a constitutive role for Sox6 in the maintenance of synaptic output from Pvalb+ neurons into adulthood.SIGNIFICANCE STATEMENT Cortical parvalbumin-expressing (Pvalb+) inhibitory neurons provide robust inhibition to neighboring pyramidal neurons, crucial for the proper functioning of cortical networks. These inhibitory neurons undergo extensive synaptic formation and maturation during the first weeks after birth and continue to dynamically maintain their synaptic output throughout adulthood. However, it remains largely unknown which transcriptional programs underlie the postnatal maturation and maintenance of Pvalb+ neurons. Here, we show that the transcription factor Sox6 cell-autonomously regulates the synaptic maintenance and output of Pvalb+ neurons until adulthood, leaving unaffected other maturational features of this neuronal population.

Conditional GWAS analysis to identify disorder-specific SNPs for psychiatric disorders.

Substantial genetic liability is shared across psychiatric disorders but less is known about risk variants that are specific to a given disorder. We used multi-trait conditional and joint analysis (mtCOJO) to adjust GWAS summary statistics of one disorder for the effects of genetically correlated traits to identify putative disorder-specific SNP associations. We applied mtCOJO to summary statistics for five psychiatric disorders from the Psychiatric Genomics Consortium-schizophrenia (SCZ), bipolar disorder (BIP), major depression (MD), attention-deficit hyperactivity disorder (ADHD) and autism (AUT). Most genome-wide significant variants for these disorders had evidence of pleiotropy (i.e., impact on multiple psychiatric disorders) and hence have reduced mtCOJO conditional effect sizes. However, subsets of genome-wide significant variants had larger conditional effect sizes consistent with disorder-specific effects: 15 of 130 genome-wide significant variants for schizophrenia, 5 of 40 for major depression, 3 of 11 for ADHD and 1 of 2 for autism. We show that decreased expression of VPS29 in the brain may increase risk to SCZ only and increased expression of CSE1L is associated with SCZ and MD, but not with BIP. Likewise, decreased expression of PCDHA7 in the brain is linked to increased risk of MD but decreased risk of SCZ and BIP.

Intrinsic electrophysiological properties predict variability in morphology and connectivity among striatal Parvalbumin-expressing Pthlh-cells.

Determining the cellular content of the nervous system in terms of cell types and the rules of their connectivity represents a fundamental challenge to the neurosciences. The recent advent of high-throughput techniques, such as single-cell RNA-sequencing has allowed for greater resolution in the identification of cell types and/or states. Although most of the current neuronal classification schemes comprise discrete clusters, several recent studies have suggested that, perhaps especially, within the striatum, neuronal populations exist in continua, with regards to both their molecular and electrophysiological properties. Whether these continua are stable properties, established during development, or if they reflect acute differences in activity-dependent regulation of critical genes is currently unknown. We set out to determine whether gradient-like molecular differences in the recently described Pthlh-expressing inhibitory interneuron population, which contains the Pvalb-expressing cells, correlate with differences in morphological and connectivity properties. We show that morphology and long-range inputs correlate with a spatially organized molecular and electrophysiological gradient of Pthlh-interneurons, suggesting that the processing of different types of information (by distinct anatomical striatal regions) has different computational requirements.

A community-based transcriptomics classification and nomenclature of neocortical cell types.

To understand the function of cortical circuits, it is necessary to catalog their cellular diversity. Past attempts to do so using anatomical, physiological or molecular features of cortical cells have not resulted in a unified taxonomy of neuronal or glial cell types, partly due to limited data. Single-cell transcriptomics is enabling, for the first time, systematic high-throughput measurements of cortical cells and generation of datasets that hold the promise of being complete, accurate and permanent. Statistical analyses of these data reveal clusters that often correspond to cell types previously defined by morphological or physiological criteria and that appear conserved across cortical areas and species. To capitalize on these new methods, we propose the adoption of a transcriptome-based taxonomy of cell types for mammalian neocortex. This classification should be hierarchical and use a standardized nomenclature. It should be based on a probabilistic definition of a cell type and incorporate data from different approaches, developmental stages and species. A community-based classification and data aggregation model, such as a knowledge graph, could provide a common foundation for the study of cortical circuits. This community-based classification, nomenclature and data aggregation could serve as an example for cell type atlases in other parts of the body.

Heterogeneous somatostatin-expressing neuron population in mouse ventral tegmental area.

The cellular architecture of the ventral tegmental area (VTA), the main hub of the brain reward system, remains only partially characterized. To extend the characterization to inhibitory neurons, we have identified three distinct subtypes of somatostatin (Sst)-expressing neurons in the mouse VTA. These neurons differ in their electrophysiological and morphological properties, anatomical localization, as well as mRNA expression profiles. Importantly, similar to cortical Sst-containing interneurons, most VTA Sst neurons express GABAergic inhibitory markers, but some of them also express glutamatergic excitatory markers and a subpopulation even express dopaminergic markers. Furthermore, only some of the proposed marker genes for cortical Sst neurons were expressed in the VTA Sst neurons. Physiologically, one of the VTA Sst neuron subtypes locally inhibited neighboring dopamine neurons. Overall, our results demonstrate the remarkable complexity and heterogeneity of VTA Sst neurons and suggest that these cells are multifunctional players in the midbrain reward circuitry.

Genetic identification of cell types underlying brain complex traits yields insights into the etiology of Parkinson's disease.

Genome-wide association studies have discovered hundreds of loci associated with complex brain disorders, but it remains unclear in which cell types these loci are active. Here we integrate genome-wide association study results with single-cell transcriptomic data from the entire mouse nervous system to systematically identify cell types underlying brain complex traits. We show that psychiatric disorders are predominantly associated with projecting excitatory and inhibitory neurons. Neurological diseases were associated with different cell types, which is consistent with other lines of evidence. Notably, Parkinson's disease was genetically associated not only with cholinergic and monoaminergic neurons (which include dopaminergic neurons) but also with enteric neurons and oligodendrocytes. Using post-mortem brain transcriptomic data, we confirmed alterations in these cells, even at the earliest stages of disease progression. Our study provides an important framework for understanding the cellular basis of complex brain maladies, and reveals an unexpected role of oligodendrocytes in Parkinson's disease.

Author Correction: Genome-wide meta-analysis identifies new loci and functional pathways influencing Alzheimer's disease risk.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Spatiotemporal mapping of RNA editing in the developing mouse brain using in situ sequencing reveals regional and cell-type-specific regulation.

BACKGROUND: Adenosine-to-inosine (A-to-I) RNA editing is a process that contributes to the diversification of proteins that has been shown to be essential for neurotransmission and other neuronal functions. However, the spatiotemporal and diversification properties of RNA editing in the brain are largely unknown. Here, we applied in situ sequencing to distinguish between edited and unedited transcripts in distinct regions of the mouse brain at four developmental stages, and investigate the diversity of the RNA landscape. RESULTS: We analyzed RNA editing at codon-altering sites using in situ sequencing at single-cell resolution, in combination with the detection of individual ADAR enzymes and specific cell type marker transcripts. This approach revealed cell-type-specific regulation of RNA editing of a set of transcripts, and developmental and regional variation in editing levels for many of the targeted sites. We found increasing editing diversity throughout development, which arises through regional- and cell type-specific regulation of ADAR enzymes and target transcripts. CONCLUSIONS: Our single-cell in situ sequencing method has proved useful to study the complex landscape of RNA editing and our results indicate that this complexity arises due to distinct mechanisms of regulating individual RNA editing sites, acting both regionally and in specific cell types.