Improving reliability and absolute quantification of human brain microarray data by filtering and scaling probes using RNA-Seq

Age-associated cortical similarity networks correlate with cell type-specific transcriptional signatures

Human brain structure shows heterogeneous patterns of change across adults aging and is associated with cognition. However, the relationship between cortical structural changes during aging and gene transcription signatures remains unclear. Here, using structural magnetic resonance imaging data of two separate cohorts of healthy participants from the Cambridge Centre for Aging and Neuroscience (n = 454, 18-87 years) and Dallas Lifespan Brain Study (n = 304, 20-89 years) and a transcriptome dataset, we investigated the link between cortical morphometric similarity network and brain-wide gene transcription. In two cohorts, we found reproducible morphometric similarity network change patterns of decreased morphological similarity with age in cognitive related areas (mainly located in superior frontal and temporal cortices), and increased morphological similarity in sensorimotor related areas (postcentral and lateral occipital cortices). Changes in morphometric similarity network showed significant spatial correlation with the expression of age-related genes that enriched to synaptic-related biological processes, synaptic abnormalities likely accounting for cognitive decline. Transcription changes in astrocytes, microglia, and neuronal cells interpreted most of the age-related morphometric similarity network changes, which suggest potential intervention and therapeutic targets for cognitive decline. Taken together, by linking gene transcription signatures to cortical morphometric similarity network, our findings might provide molecular and cellular substrates for cortical structural changes related to cognitive decline across adults aging.

Data Mining of Microarray Datasets in Translational Neuroscience

Data mining involves the computational analysis of a plethora of publicly available datasets to generate new hypotheses that can be further validated by experiments for the improved understanding of the pathogenesis of neurodegenerative diseases. Although the number of sequencing datasets is on the rise, microarray analysis conducted on diverse biological samples represent a large collection of datasets with multiple web-based programs that enable efficient and convenient data analysis. In this review, we first discuss the selection of biological samples associated with neurological disorders, and the possibility of a combination of datasets, from various types of samples, to conduct an integrated analysis in order to achieve a holistic understanding of the alterations in the examined biological system. We then summarize key approaches and studies that have made use of the data mining of microarray datasets to obtain insights into translational neuroscience applications, including biomarker discovery, therapeutic development, and the elucidation of the pathogenic mechanisms of neurodegenerative diseases. We further discuss the gap to be bridged between microarray and sequencing studies to improve the utilization and combination of different types of datasets, together with experimental validation, for more comprehensive analyses. We conclude by providing future perspectives on integrating multi-omics, to advance precision phenotyping and personalized medicine for neurodegenerative diseases.

Microstructural imaging and transcriptomics of the basal forebrain in first-episode psychosis

Cholinergic dysfunction has been implicated in the pathophysiology of psychosis and psychiatric disorders such as schizophrenia, depression, and bipolar disorder. The basal forebrain (BF) cholinergic nuclei, defined as cholinergic cell groups Ch1-3 and Ch4 (Nucleus Basalis of Meynert; NBM), provide extensive cholinergic projections to the rest of the brain. Here, we examined microstructural neuroimaging measures of the cholinergic nuclei in patients with untreated psychosis (~31 weeks of psychosis, <2 defined daily dose of antipsychotics) and used magnetic resonance spectroscopy (MRS) and transcriptomic data to support our findings. We used a cytoarchitectonic atlas of the BF to map the nuclei and obtained measures of myelin (quantitative T1, or qT1 as myelin surrogate) and microstructure (axial diffusion; AxD). In a clinical sample (n = 85; 29 healthy controls, 56 first-episode psychosis), we found significant correlations between qT1 of Ch1-3, left NBM and MRS-based dorsal anterior cingulate choline in healthy controls while this relationship was disrupted in FEP (p > 0.05). Case-control differences in qT1 and AxD were observed in the Ch1-3, with increased qT1 (reflecting reduced myelin content) and AxD (reflecting reduced axonal integrity). We found clinical correlates between left NBM qT1 with manic symptom severity, and AxD with negative symptom burden in FEP. Intracortical and subcortical myelin maps were derived and correlated with BF myelin. BF-cortical and BF-subcortical myelin correlations demonstrate known projection patterns from the BF. Using data from the Allen Human Brain Atlas, cholinergic nuclei showed significant enrichment for schizophrenia and depression-related genes. Cell-type specific enrichment indicated enrichment for cholinergic neuron markers as expected. Further relating the neuroimaging correlations to transcriptomics demonstrated links with cholinergic receptor genes and cell type markers of oligodendrocytes and cholinergic neurons, providing biological validity to the measures. These results provide genetic, neuroimaging, and clinical evidence for cholinergic dysfunction in schizophrenia.

Transcriptome-based polygenic score links depression-related corticolimbic gene expression changes to sex-specific brain morphology and depression risk

Studies in post-mortem human brain tissue have associated major depressive disorder (MDD) with cortical transcriptomic changes, whose potential in vivo impact remains unexplored. To address this translational gap, we recently developed a transcriptome-based polygenic risk score (T-PRS) based on common functional variants capturing 'depression-like' shifts in cortical gene expression. Here, we used a non-clinical sample of young adults (n = 482, Duke Neurogenetics Study: 53% women; aged 19.8 ± 1.2 years) to map T-PRS onto brain morphology measures, including Freesurfer-derived subcortical volume, cortical thickness, surface area, and local gyrification index, as well as broad MDD risk, indexed by self-reported family history of depression. We conducted side-by-side comparisons with a PRS independently derived from a Psychiatric Genomics Consortium (PGC) MDD GWAS (PGC-PRS), and sought to link T-PRS with diagnosis and symptom severity directly in PGC-MDD participants (n = 29,340, 59% women; 12,923 MDD cases, 16,417 controls). T-PRS was associated with smaller amygdala volume in women (t = -3.478, p = 0.001) and lower prefrontal gyrification across sexes. In men, T-PRS was associated with hypergyrification in temporal and occipital regions. Prefrontal hypogyrification mediated a male-specific indirect link between T-PRS and familial depression (b = 0.005, p = 0.029). PGC-PRS was similarly associated with lower amygdala volume and cortical gyrification; however, both effects were male-specific and hypogyrification emerged in distinct parietal and temporo-occipital regions, unassociated with familial depression. In PGC-MDD, T-PRS did not predict diagnosis (OR = 1.007, 95% CI = [0.997-1.018]) but correlated with symptom severity in men (rho = 0.175, p = 7.957 × 10^-4) in one cohort (N = 762, 48% men). Depression-like shifts in cortical gene expression have sex-specific effects on brain morphology and may contribute to broad depression vulnerability in men.

Mapping gene transcription and neurocognition across human neocortex

Regulation of gene expression drives protein interactions that govern synaptic wiring and neuronal activity. The resulting coordinated activity among neuronal populations supports complex psychological processes, yet how gene expression shapes cognition and emotion remains unknown. Here, we directly bridge the microscale and macroscale by mapping gene expression patterns to functional activation patterns across the cortical sheet. Applying unsupervised learning to the Allen Human Brain Atlas and Neurosynth databases, we identify a ventromedial-dorsolateral gradient of gene assemblies that separate affective and perceptual domains. This topographic molecular-psychological signature reflects the hierarchical organization of the neocortex, including systematic variations in cell type, myeloarchitecture, laminar differentiation and intrinsic network affiliation. In addition, this molecular-psychological signature strengthens over neurodevelopment and can be replicated in two independent repositories. Collectively, our results reveal spatially covarying transcriptomic and cognitive architectures, highlighting the influence that molecular mechanisms exert on psychological processes.

Genetic influences on hub connectivity of the human connectome

Brain network hubs are both highly connected and highly inter-connected, forming a critical communication backbone for coherent neural dynamics. The mechanisms driving this organization are poorly understood. Using diffusion-weighted magnetic resonance imaging in twins, we identify a major role for genes, showing that they preferentially influence connectivity strength between network hubs of the human connectome. Using transcriptomic atlas data, we show that connected hubs demonstrate tight coupling of transcriptional activity related to metabolic and cytoarchitectonic similarity. Finally, comparing over thirteen generative models of network growth, we show that purely stochastic processes cannot explain the precise wiring patterns of hubs, and that model performance can be improved by incorporating genetic constraints. Our findings indicate that genes play a strong and preferential role in shaping the functionally valuable, metabolically costly connections between connectome hubs.

Dynamical consequences of regional heterogeneity in the brain's transcriptional landscape

Brain regions vary in their molecular and cellular composition, but how this heterogeneity shapes neuronal dynamics is unclear. Here, we investigate the dynamical consequences of regional heterogeneity using a biophysical model of whole-brain functional magnetic resonance imaging (MRI) dynamics in humans. We show that models in which transcriptional variations in excitatory and inhibitory receptor (E:I) gene expression constrain regional heterogeneity more accurately reproduce the spatiotemporal structure of empirical functional connectivity estimates than do models constrained by global gene expression profiles or MRI-derived estimates of myeloarchitecture. We further show that regional transcriptional heterogeneity is essential for yielding both ignition-like dynamics, which are thought to support conscious processing, and a wide variance of regional-activity time scales, which supports a broad dynamical range. We thus identify a key role for E:I heterogeneity in generating complex neuronal dynamics and demonstrate the viability of using transcriptomic data to constrain models of large-scale brain function.

Characterization of Human Genes Modulated by Porphyromonas gingivalis Highlights the Ribosome, Hypothalamus, and Cholinergic Neurons

Porphyromonas gingivalis, a bacterium associated with periodontal disease, is a suspected cause of Alzheimer's disease. This bacterium is reliant on gingipain proteases, which cleave host proteins after arginine and lysine residues. To characterize gingipain susceptibility, we performed enrichment analyses of arginine and lysine proportion proteome-wide. Genes differentially expressed in brain samples with detected P. gingivalis reads were also examined. Genes from these analyses were tested for functional enrichment and specific neuroanatomical expression patterns. Proteins in the SRP-dependent cotranslational protein targeting to membrane pathway were enriched for these residues and previously associated with periodontal and Alzheimer's disease. These ribosomal genes are up-regulated in prefrontal cortex samples with detected P. gingivalis sequences. Other differentially expressed genes have been previously associated with dementia (ITM2B, MAPT, ZNF267, and DHX37). For an anatomical perspective, we characterized the expression of the P. gingivalis associated genes in the mouse and human brain. This analysis highlighted the hypothalamus, cholinergic neurons, and the basal forebrain. Our results suggest markers of neural P. gingivalis infection and link the cholinergic and gingipain hypotheses of Alzheimer's disease.

Downregulation of glial genes involved in synaptic function mitigates Huntington's disease pathogenesis

Most research on neurodegenerative diseases has focused on neurons, yet glia help form and maintain the synapses whose loss is so prominent in these conditions. To investigate the contributions of glia to Huntington's disease (HD), we profiled the gene expression alterations of Drosophila expressing human mutant Huntingtin (mHTT) in either glia or neurons and compared these changes to what is observed in HD human and HD mice striata. A large portion of conserved genes are concordantly dysregulated across the three species; we tested these genes in a high-throughput behavioral assay and found that downregulation of genes involved in synapse assembly mitigated pathogenesis and behavioral deficits. To our surprise, reducing dNRXN3 function in glia was sufficient to improve the phenotype of flies expressing mHTT in neurons, suggesting that mHTT's toxic effects in glia ramify throughout the brain. This supports a model in which dampening synaptic function is protective because it attenuates the excitotoxicity that characterizes HD.

When a neuron dies, through injury or disease, the body loses all communication that passes through it. The brain compensates by rerouting the flow of information through other neurons in the network. Eventually, if the loss of neurons becomes too great, compensation becomes impossible. This process happens in Alzheimer's, Parkinson's, and Huntington's disease. In the case of Huntington's disease, the cause is mutation to a single gene known as huntingtin. The mutation is present in every cell in the body but causes particular damage to parts of the brain involved in mood, thinking and movement.

Neurons and other cells respond to mutations in the huntingtin gene by turning the activities of other genes up or down, but it is not clear whether all of these changes contribute to the damage seen in Huntington's disease. In fact, it is possible that some of the changes are a result of the brain trying to protect itself. So far, most research on this subject has focused on neurons because the huntingtin gene plays a role in maintaining healthy neuronal connections. But, given that all cells carry the mutated gene, it is likely that other cells are also involved. The glia are a diverse group of cells that support the brain, providing care and sustenance to neurons. These cells have a known role in maintaining the connections between neurons and may also have play a role in either causing or correcting the damage seen in Huntington's disease.

The aim of Onur et al. was to find out which genes are affected by having a mutant huntingtin gene in neurons or glia, and whether severity of Huntington’s disease improved or worsened when the activity of these genes changed. First, Onur et al. identified genes affected by mutant huntingtin by comparing healthy human brains to the brains of people with Huntington's disease. Repeating the same comparison in mice and fruit flies identified genes affected in the same way across all three species, revealing that, in Huntington's disease, the brain dials down glial cell genes involved in maintaining neuronal connections.

To find out how these changes in gene activity affect disease severity and progression, Onur et al. manipulated the activity of each of the genes they had identified in fruit flies that carried mutant versions of huntingtin either in neurons, in glial cells or in both cell types. They then filmed the flies to see the effects of the manipulation on movement behaviors, which are affected by Huntington’s disease. This revealed that purposely lowering the activity of the glial genes involved in maintaining connections between neurons improved the symptoms of the disease, but only in flies who had mutant huntingtin in their glial cells. This indicates that the drop in activity of these genes observed in Huntington’s disease is the brain trying to protect itself.

This work suggests that it is important to include glial cells in studies of neurological disorders. It also highlights the fact that changes in gene expression as a result of a disease are not always bad. Many alterations are compensatory, and try to either make up for or protect cells affected by the disease. Therefore, it may be important to consider whether drugs designed to treat a condition by changing levels of gene activity might undo some of the body's natural protection. Working out which changes drive disease and which changes are protective will be essential for designing effective treatments.

Donor-Specific Transcriptomic Analysis of Alzheimer's Disease-Associated Hypometabolism Highlights a Unique Donor, Ribosomal Proteins and Microglia

Alzheimer's disease (AD) starts decades before clinical symptoms appear. Low-glucose utilization in regions of the cerebral cortex marks early AD. To identify these regions, we conducted a voxel-wise meta-analysis of previous studies conducted with positron emission tomography that compared AD patients with healthy controls. The resulting map marks hypometabolism in the posterior cingulate, middle frontal, angular gyrus, and middle and inferior temporal regions. Using the Allen Human Brain Atlas, we identified genes that show spatial correlation across the cerebral cortex between their expression and this hypometabolism. Of the six brains in the Atlas, one demonstrated a strong spatial correlation between gene expression and hypometabolism. Previous neuropathological assessment of this brain from a 39-year-old male noted a neurofibrillary tangle in the entorhinal cortex. Using the transcriptomic data, we estimate lower proportions of neurons and more microglia in the hypometabolic regions when comparing this donor's brain with the other five donors. Within this single brain, signal recognition particle (SRP)-dependent cotranslational protein targeting genes, which encode primarily cytosolic ribosome proteins, are highly expressed in the hypometabolic regions. Analyses of human and mouse data show that expression of these genes increases progressively across AD-associated states of microglial activation. In addition, genes involved in cell killing, chronic inflammation, ubiquitination, tRNA aminoacylation, and vacuole sorting are associated with the hypometabolism map. These genes suggest disruption of the protein life cycle and neuroimmune activation. Taken together, our molecular characterization reveals a link to AD-associated hypometabolism that may be relevant to preclinical stages of AD.

Comparison Between Expression Microarrays and RNA-Sequencing Using UKBEC Dataset Identified a trans-eQTL Associated with MPZ Gene in Substantia Nigra

In recent years, the advantages of RNA-sequencing (RNA-Seq) have made it the platform of choice for measuring gene expression over traditional microarrays. However, RNA-Seq comes with bioinformatical challenges and higher computational costs. Therefore, this study set out to assess whether the increased depth of transcriptomic information facilitated by RNA-Seq is worth the increased computation over microarrays, specifically at three levels: absolute expression levels, differentially expressed genes identification, and expression QTL (eQTL) mapping in regions of the human brain. Using the United Kingdom Brain Expression Consortium (UKBEC) dataset, there is high agreement of gene expression levels measured by microarrays and RNA-seq when quantifying absolute expression levels and when identifying differentially expressed genes. These findings suggest that depending on the aims of a study, the relative ease of working with microarray data may outweigh the computational time and costs of RNA-Seq pipelines. On the other, there was low agreement when mapping eQTLs. However, a number of eQTLs associated with genes that play important roles in the brain were found in both platforms. For example, a trans-eQTL was mapped that is associated with the MPZ gene in the substantia nigra. These eQTLs that we have highlighted are extremely promising candidates that merit further investigation.

Development of an RNA sequencing-based prognostic gene signature in multiple myeloma

Several prognostic gene signatures have been developed to predict the clinical outcome in patients with multiple myeloma (MM). The most salient disadvantage of the previous signatures is their non-reproducibility in external datasets. Given the disadvantages and the superiority of RNA sequencing over microarrays in transcriptome profiling to produce more reliable outputs, we sought to develop a reproducible RNA sequencing-based prognostic gene signature for MM. Genes significantly associated with survival were detected in The Cancer Genome Atlas (TCGA) MM RNA sequencing dataset (MMRF-CoMMpass) (n = 412) through a strict pipeline containing four rigid filters. The reproducibility of the selected genes was checked in an independent dataset (GSE24080), containing 559 newly diagnosed patients with MM. The RNA sequencing-based prognostic signature was reconstructed based on the final genes in the training dataset (MMRF-CoMMpass) and externally validated in five independent datasets (i.e. GSE2658, GSE13624, GSE9782, GSE6477 and GSE57317), containing 1461 MM cases. The RNA sequencing-based signature was reconstructed using finally five reproducible genes: CCT2, CKS1B, PRKDC, NONO and UBE2A. This signature was able to robustly discriminate between low- and high-risk patients in both training and validation datasets (Ps ≤ 0·001). Our signature was also independent of and more powerful than the routine MM prognostic factors (i.e. β2-microglobulin, albumin, age and sex) (Ps ≤ 0·01). Treatment regimens had no effect on RNA sequencing-based signature insofar as this signature succeeded in predicting the clinical outcome in various treatment groups (Ps ≤ 0·001).

It is time to apply biclustering: a comprehensive review of biclustering applications in biological and biomedical data

Biclustering is a powerful data mining technique that allows clustering of rows and columns, simultaneously, in a matrix-format data set. It was first applied to gene expression data in 2000, aiming to identify co-expressed genes under a subset of all the conditions/samples. During the past 17 years, tens of biclustering algorithms and tools have been developed to enhance the ability to make sense out of large data sets generated in the wake of high-throughput omics technologies. These algorithms and tools have been applied to a wide variety of data types, including but not limited to, genomes, transcriptomes, exomes, epigenomes, phenomes and pharmacogenomes. However, there is still a considerable gap between biclustering methodology development and comprehensive data interpretation, mainly because of the lack of knowledge for the selection of appropriate biclustering tools and further supporting computational techniques in specific studies. Here, we first deliver a brief introduction to the existing biclustering algorithms and tools in public domain, and then systematically summarize the basic applications of biclustering for biological data and more advanced applications of biclustering for biomedical data. This review will assist researchers to effectively analyze their big data and generate valuable biological knowledge and novel insights with higher efficiency.

Interpretation of differential gene expression results of RNA-seq data: review and integration

Differential gene expression (DGE) analysis is one of the most common applications of RNA-sequencing (RNA-seq) data. This process allows for the elucidation of differentially expressed genes across two or more conditions and is widely used in many applications of RNA-seq data analysis. Interpretation of the DGE results can be nonintuitive and time consuming due to the variety of formats based on the tool of choice and the numerous pieces of information provided in these results files. Here we reviewed DGE results analysis from a functional point of view for various visualizations. We also provide an R/Bioconductor package, Visualization of Differential Gene Expression Results using R, which generates information-rich visualizations for the interpretation of DGE results from three widely used tools, Cuffdiff, DESeq2 and edgeR. The implemented functions are also tested on five real-world data sets, consisting of one human, one Malus domestica and three Vitis riparia data sets.

Transcriptomic Characterization of the Human Insular Cortex and Claustrum

The insular cortex has been linked to a multitude of functions. In contrast, the nearby claustrum is a densely connected subcortical region with unclear function. To view the insula-claustrum region from the molecular perspective we analyzed the transcriptomic profile of these areas in six adult and four fetal human brains. We identified marker genes with specific expression and performed transcriptome-wide tests for enrichment of biological processes, molecular functions, and cellular components. In addition, specific insular and claustral expression of genes pertaining to diseases, addiction, and depression was tested. At the anatomical level, we used brain-wide analyses to determine the specificity of our results and to determine the transcriptomic similarity of the insula-claustrum region. We found UCMA to be the most significantly enriched gene in the insular cortex and confirmed specific expression of NR4A2, NTNG2, and LXN in the claustrum. Furthermore, the insula was found to have enriched expression of genes associated with mood disorders, learning, cardiac muscle contraction, oxygen transport, glutamate and dopamine signaling. Specific expression in the claustrum was enriched for genes pertaining to human immunodeficiency virus (HIV), severe intellectual disability, epileptic encephalopathy, intracellular transport, spine development, and macroautophagy. We tested for enrichment of genes related to addiction and depression, but they were generally not highly specific to the insula-claustrum region. Exceptions include high insular expression of genes linked to cocaine abuse and genes associated with ever smoking in the claustrum. Brain-wide, we find that markers of the adult claustrum are most specifically expressed in the fetal and adult insula. Altogether, our results provide a novel molecular perspective on the unique properties of the insula and claustrum.

Investigation of optimal pathways for preeclampsia using network-based guilt by association algorithm

This study investigated optimal pathways for preeclampsia (PE) utilizing the network-based guilt by association (GBA) algorithm. The inference method consisted of four steps: preparing differentially expressed genes (DEGs) between PE patients and normal controls from gene expression data; constructing co-expression network (CEN) for DEGs utilizing Spearman's correlation coefficient (SCC) method; and predicting optimal pathways by network-based GBA algorithm of which the area under the receiver operating characteristics curve (AUROC) was gained for each pathway. There were 351 DEGs and 61,425 edges in the CEN for PE. Subsequently, 53 pathways were obtained with a good classification performance (AUROC >0.5). AUROC for 9 was >0.9 and defined as optimal pathways, especially microRNAs in cancer (AUROC=0.9966), gap junction (AUROC=0.9922), and pathogenic Escherichia coli infection (AUROC=0.9888). Nine optimal pathways were identified through comprehensive analysis of data from PE patients, which might shed new light on uncovering molecular and pathological mechanism of PE.

A practical guide to linking brain-wide gene expression and neuroimaging data

The recent availability of comprehensive, brain-wide gene expression atlases such as the Allen Human Brain Atlas (AHBA) has opened new opportunities for understanding how spatial variations on molecular scale relate to the macroscopic neuroimaging phenotypes. A rapidly growing body of literature is demonstrating relationships between gene expression and diverse properties of brain structure and function, but approaches for combining expression atlas data with neuroimaging are highly inconsistent, with substantial variations in how the expression data are processed. The degree to which these methodological variations affect findings is unclear. Here, we outline a seven-step analysis pipeline for relating brain-wide transcriptomic and neuroimaging data and compare how different processing choices influence the resulting data. We suggest that studies using the AHBA should work towards a unified data processing pipeline to ensure consistent and reproducible results in this burgeoning field.

Transcriptomic Characterization of the Human Habenula Highlights Drug Metabolism and the Neuroimmune System

Due to size and accessibility, most information about the habenula is derived from rodent studies. To better understand the molecular signature of the habenula we characterized the genes that have high expression in the habenula. We compared anatomical expression profiles of three normal adult human brains and four fetal brains. We used gene set enrichment analyses to determine if genes annotated to specific molecular functions, cellular components, and biological processes are enriched in the habenula. We also tested gene sets related to depression and addiction to determine if they uniquely involve the habenula. As expected, we observed high habenular expression of GPR151, nicotinic cholinergic receptors, and cilia-associated genes (medial division). Genes identified in genetic studies of smoking and associated with nicotine response were enriched in the habenula. Genes associated with major depressive disorder did not have enriched expression in the habenula but genes negatively correlated with hedonic well-being were, providing a link to anhedonia. We observed enrichment of genes associated with diseases that are comorbid with addictions (hematopoiesis, thrombosis, liver cirrhosis, pneumonia, and pulmonary fibrosis) and depression (rheumatoid arthritis, multiple sclerosis, and kidney disease). These inflammatory diseases mark a neuroimmune signature that is supported by genes associated with mast cells, acute inflammatory response, and leukocyte migration. We also found enrichment of cytochrome p450 genes suggesting the habenula is uniquely sensitive to endogenous and xenobiotic compounds. Our results suggest the habenula receives negative reward signals from immune and drug processing molecules. This is consistent with the habenular role in the "anti-reward" system and suggests it may be a key bridge between autoimmune disorders, drug use, and psychiatric diseases.

A New Machine Learning-Based Framework for Mapping Uncertainty Analysis in RNA-Seq Read Alignment and Gene Expression Estimation

One of the main benefits of using modern RNA-Sequencing (RNA-Seq) technology is the more accurate gene expression estimations compared with previous generations of expression data, such as the microarray. However, numerous issues can result in the possibility that an RNA-Seq read can be mapped to multiple locations on the reference genome with the same alignment scores, which occurs in plant, animal, and metagenome samples. Such a read is so-called a multiple-mapping read (MMR). The impact of these MMRs is reflected in gene expression estimation and all downstream analyses, including differential gene expression, functional enrichment, etc. Current analysis pipelines lack the tools to effectively test the reliability of gene expression estimations, thus are incapable of ensuring the validity of all downstream analyses. Our investigation into 95 RNA-Seq datasets from seven plant and animal species (totaling 1,951 GB) indicates an average of roughly 22% of all reads are MMRs. Here we present a machine learning-based tool called GeneQC (Gene expression Quality Control), which can accurately estimate the reliability of each gene's expression level derived from an RNA-Seq dataset. The underlying algorithm is designed based on extracted genomic and transcriptomic features, which are then combined using elastic-net regularization and mixture model fitting to provide a clearer picture of mapping uncertainty for each gene. GeneQC allows researchers to determine reliable expression estimations and conduct further analysis on the gene expression that is of sufficient quality. This tool also enables researchers to investigate continued re-alignment methods to determine more accurate gene expression estimates for those with low reliability. Application of GeneQC reveals high level of mapping uncertainty in plant samples and limited, severe mapping uncertainty in animal samples. GeneQC is freely available at http://bmbl.sdstate.edu/GeneQC/home.html.

An anatomic transcriptional atlas of human glioblastoma

Glioblastoma is an aggressive brain tumor that carries a poor prognosis. The tumor's molecular and cellular landscapes are complex, and their relationships to histologic features routinely used for diagnosis are unclear. We present the Ivy Glioblastoma Atlas, an anatomically based transcriptional atlas of human glioblastoma that aligns individual histologic features with genomic alterations and gene expression patterns, thus assigning molecular information to the most important morphologic hallmarks of the tumor. The atlas and its clinical and genomic database are freely accessible online data resources that will serve as a valuable platform for future investigations of glioblastoma pathogenesis, diagnosis, and treatment.

Transcriptomic characterization of MRI contrast with focus on the T1-w/T2-w ratio in the cerebral cortex

Magnetic resonance (MR) images of the brain are of immense clinical and research utility. At the atomic and subatomic levels, the sources of MR signals are well understood. However, we lack a comprehensive understanding of the macromolecular correlates of MR signal contrast. To address this gap, we used genome-wide measurements to correlate gene expression with MR signal intensity across the cerebral cortex in the Allen Human Brain Atlas (AHBA). We focused on the ratio of T1-weighted and T2-weighted intensities (T1-w/T2-w ratio image), which is considered to be a useful proxy for myelin content. As expected, we found enrichment of positive correlations between myelin-associated genes and the ratio image, supporting its use as a myelin marker. Genome-wide, there was an association with protein mass, with genes coding for heavier proteins expressed in regions with high T1-w/T2-w values. Oligodendrocyte gene markers were strongly correlated with the T1-w/T2-w ratio, but this was not driven by myelin-associated genes. Mitochondrial genes exhibit the strongest relationship, showing higher expression in regions with low T1-w/T2-w ratio. This may be due to the pH gradient in mitochondria as genes up-regulated by pH in the brain were also highly correlated with the ratio. While we corroborate associations with myelin and synaptic plasticity, differences in the T1-w/T2-w ratio across the cortex are more strongly linked to molecule size, oligodendrocyte markers, mitochondria, and pH. We evaluate correlations between AHBA transcriptomic measurements and a group averaged T1-w/T2-w ratio image, showing agreement with in-sample results. Expanding our analysis to the whole brain results in strong positive T1-w/T2-w correlations for immune system, inflammatory disease, and microglia marker genes. Genes with negative correlations were enriched for neuron markers and synaptic plasticity genes. Lastly, our findings are similar when performed on T1-w or inverted T2-w intensities alone. These results provide a molecular characterization of MR contrast that will aid interpretation of future MR studies of the brain.

Complement Gene Expression Correlates with Superior Frontal Cortical Thickness in Humans

Recent work suggests that genes encoding complement proteins that are active in the innate immune system may confer risk for schizophrenia by disrupting typical synaptic pruning in late adolescence. Alterations in the complement pathway may contribute to aberrant cortical thinning in schizophrenia prodromes and reduced prefrontal cortical thickness in chronic schizophrenia patients; however, this theory needs to be translated to humans. We conducted a series of analyses in a sample of adult Swedish twins enriched for schizophrenia (N=129) to assess the plausibility of a relationship between complement gene expression and cortical thickness that could go awry in the etiology of schizophrenia. First, we identified that peripheral mRNA expression levels of two complement genes (C5, SERPING1) made unique contributions to the variance in superior frontal cortical thickness among all participants. Vertex-wise maps of the association between gene expression levels and thickness across the cortex suggested that this relationship was especially strong with SERPING1 in the superior frontal region, consistent with the pattern of disruption in cortical thickness observed in schizophrenia. Additional analyses identified that these genes are expressed in the human superior frontal cortex, that heritable genetic factors influence SERPING1 gene expression levels, and that these associations are observed regardless of case status. These findings provide initial evidence linking the complement system with cortical thinning in humans, a process potentially involved in the pathogenesis of schizophrenia.

RNA-Seq and Expression Arrays: Selection Guidelines for Genome-Wide Expression Profiling

The development of genome-wide gene expression profiling technologies over the past two decades has produced great opportunity for researchers to explore the transcriptome and to better understand biological systems and their perturbation. In this chapter we provide an overview of microarray and massively parallel sequencing technologies and their application to gene expression analysis. We discuss factors that impact expression data generation and analysis that which should be considered in the application of these technology platforms. We further present the results of a simple illustration study to highlight performance similarities and differences in expression profiling of protein-coding mRNAs with each platform. Based on technical and analytical differences between the two platforms, reports in the literature comparing arrays and RNA-Seq for gene expression, and our own example study and experience, we provide recommendations for platform selection for gene expression studies.

Oxytocin under opioid antagonism leads to supralinear enhancement of social attention

To provide new preclinical evidence toward improving the efficacy of oxytocin (OT) in treating social dysfunction, we tested the benefit of administering OT under simultaneously induced opioid antagonism during dyadic gaze interactions in monkeys. OT coadministered with a μ-opioid receptor antagonist, naloxone, invoked a supralinear enhancement of prolonged and selective social attention, producing a stronger effect than the summed effects of each administered separately. These effects were consistently observed when averaging over entire sessions, as well as specifically following events of particular social importance, including mutual eye contact and mutual reward receipt. Furthermore, attention to various facial regions was differentially modulated depending on social context. Using the Allen Institute's transcriptional atlas, we further established the colocalization of μ-opioid and κ-opioid receptor genes and OT genes at the OT-releasing sites in the human brain. These data across monkeys and humans support a regulatory relationship between the OT and opioid systems and suggest that administering OT under opioid antagonism may boost the therapeutic efficacy of OT for enhancing social cognition.

Brain transcriptome atlases: a computational perspective

The immense complexity of the mammalian brain is largely reflected in the underlying molecular signatures of its billions of cells. Brain transcriptome atlases provide valuable insights into gene expression patterns across different brain areas throughout the course of development. Such atlases allow researchers to probe the molecular mechanisms which define neuronal identities, neuroanatomy, and patterns of connectivity. Despite the immense effort put into generating such atlases, to answer fundamental questions in neuroscience, an even greater effort is needed to develop methods to probe the resulting high-dimensional multivariate data. We provide a comprehensive overview of the various computational methods used to analyze brain transcriptome atlases.

Unveiling gene trait relationship by cross-platform meta-analysis on Chinese hamster ovary cell transcriptome

In the past few years, transcriptome analysis has been increasingly employed to better understand the physiology of Chinese hamster ovary (CHO) cells at a global level. As more transcriptome data accumulated, meta-analysis on data sets collected from various sources can potentially provide better insights on common properties of those cells. Here, we performed meta-analysis on transcriptome data of different CHO cell lines obtained using NimbleGen or Affymetrix microarray platforms. Hierarchical clustering, non-negative matrix factorization (NMF) analysis, and principal component analysis (PCA) accordantly showed the samples were clustered into two groups: one consists of adherent cells in serum-containing medium, and the other suspension cells in serum-free medium. Genes that were differentially expressed between the two clusters were enriched in a few functional classes by Database for Annotation, Visualization, and Integrated Discovery (DAVID) of which many were common with the enriched gene sets identified by Gene Set Enrichment Analysis (GSEA), including extracellular matrix (ECM) receptor interaction, cell adhesion molecules (CAMs), and lipid related metabolism pathways. Despite the heterogeneous sources of the cell samples, the adherent and suspension growth characteristics and serum-supplementation appear to be a dominant feature in the transcriptome. The results demonstrated that meta-analysis of transcriptome could uncover features in combined data sets that individual data set might not reveal. As transcriptome data sets accumulate over time, meta-analysis will become even more revealing. Biotechnol. Bioeng. 2017;114: 1583-1592. © 2017 Wiley Periodicals, Inc.

A comprehensive transcriptional map of primate brain development

The transcriptional underpinnings of brain development remain poorly understood, particularly in humans and closely related non-human primates. We describe a high-resolution transcriptional atlas of rhesus monkey (Macaca mulatta) brain development that combines dense temporal sampling of prenatal and postnatal periods with fine anatomical division of cortical and subcortical regions associated with human neuropsychiatric disease. Gene expression changes more rapidly before birth, both in progenitor cells and maturing neurons. Cortical layers and areas acquire adult-like molecular profiles surprisingly late in postnatal development. Disparate cell populations exhibit distinct developmental timing of gene expression, but also unexpected synchrony of processes underlying neural circuit construction including cell projection and adhesion. Candidate risk genes for neurodevelopmental disorders including primary microcephaly, autism spectrum disorder, intellectual disability, and schizophrenia show disease-specific spatiotemporal enrichment within developing neocortex. Human developmental expression trajectories are more similar to monkey than rodent, although approximately 9% of genes show human-specific regulation with evidence for prolonged maturation or neoteny compared to monkey.

A Quantitative System-Scale Characterization of the Metabolism of Clostridium acetobutylicum

Engineering industrial microorganisms for ambitious applications, for example, the production of second-generation biofuels such as butanol, is impeded by a lack of knowledge of primary metabolism and its regulation. A quantitative system-scale analysis was applied to the biofuel-producing bacterium Clostridium acetobutylicum, a microorganism used for the industrial production of solvent. An improved genome-scale model, iCac967, was first developed based on thorough biochemical characterizations of 15 key metabolic enzymes and on extensive literature analysis to acquire accurate fluxomic data. In parallel, quantitative transcriptomic and proteomic analyses were performed to assess the number of mRNA molecules per cell for all genes under acidogenic, solventogenic, and alcohologenic steady-state conditions as well as the number of cytosolic protein molecules per cell for approximately 700 genes under at least one of the three steady-state conditions. A complete fluxomic, transcriptomic, and proteomic analysis applied to different metabolic states allowed us to better understand the regulation of primary metabolism. Moreover, this analysis enabled the functional characterization of numerous enzymes involved in primary metabolism, including (i) the enzymes involved in the two different butanol pathways and their cofactor specificities, (ii) the primary hydrogenase and its redox partner, (iii) the major butyryl coenzyme A (butyryl-CoA) dehydrogenase, and (iv) the major glyceraldehyde-3-phosphate dehydrogenase. This study provides important information for further metabolic engineering of C. acetobutylicum to develop a commercial process for the production of n-butanol.

Currently, there is a resurgence of interest in Clostridium acetobutylicum, the biocatalyst of the historical Weizmann process, to produce n-butanol for use both as a bulk chemical and as a renewable alternative transportation fuel. To develop a commercial process for the production of n-butanol via a metabolic engineering approach, it is necessary to better characterize both the primary metabolism of C. acetobutylicum and its regulation. Here, we apply a quantitative system-scale analysis to acidogenic, solventogenic, and alcohologenic steady-state C. acetobutylicum cells and report for the first time quantitative transcriptomic, proteomic, and fluxomic data. This approach allows for a better understanding of the regulation of primary metabolism and for the functional characterization of numerous enzymes involved in primary metabolism.

Multi-platform assessment of transcriptional profiling technologies utilizing a precise probe mapping methodology

Background

The arrival of RNA-seq as a high-throughput method competitive to the established microarray technologies has necessarily driven a need for comparative evaluation. To date, cross-platform comparisons of these technologies have been relatively few in number of platforms analyzed and were typically gene name annotation oriented. Here, we present a more extensive and yet precise assessment to elucidate differences and similarities in performance of numerous aspects including dynamic range, fidelity of raw signal and fold-change with sample titration, and concordance with qRT-PCR (TaqMan). To ensure that these results were not confounded by incompatible comparisons, we introduce the concept of probe mapping directed “transcript pattern”. A transcript pattern identifies probe(set)s across platforms that target a common set of transcripts for a specific gene. Thus, three levels of data were examined: entire data sets, data derived from a subset of 15,442 RefSeq genes common across platforms, and data derived from the transcript pattern defined subset of 7,034 RefSeq genes.

Results

In general, there were substantial core similarities between all 6 platforms evaluated; but, to varying degrees, the two RNA-seq protocols outperformed three of the four microarray platforms in most categories. Notably, a fourth microarray platform, Agilent with a modified protocol, was comparable, or marginally superior, to the RNA-seq protocols within these same assessments, especially in regards to fold-change evaluation. Furthermore, these 3 platforms (Agilent and two RNA-seq methods) demonstrated over 80 % fold-change concordance with the gold standard qRT-PCR (TaqMan).

Conclusions

This study suggests that microarrays can perform on nearly equal footing with RNA-seq, in certain key features, specifically when the dynamic range is comparable. Furthermore, the concept of a transcript pattern has been introduced that may minimize potential confounding factors of multi-platform comparison and may be useful for similar evaluations.

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-015-1913-6) contains supplementary material, which is available to authorized users.

Seq-ing improved gene expression estimates from microarrays using machine learning

BACKGROUND

Quantifying gene expression by RNA-Seq has several advantages over microarrays, including greater dynamic range and gene expression estimates on an absolute, rather than a relative scale. Nevertheless, microarrays remain in widespread use, demonstrated by the ever-growing numbers of samples deposited in public repositories.

RESULTS

We propose a novel approach to microarray analysis that attains many of the advantages of RNA-Seq. This method, called Machine Learning of Transcript Expression (MaLTE), leverages samples for which both microarray and RNA-Seq data are available, using a Random Forest to learn the relationship between the fluorescence intensity of sets of microarray probes and RNA-Seq transcript expression estimates. We trained MaLTE on data from the Genotype-Tissue Expression (GTEx) project, consisting of Affymetrix gene arrays and RNA-Seq from over 700 samples across a broad range of human tissues.

CONCLUSION

This approach can be used to accurately estimate absolute expression levels from microarray data, at both gene and transcript level, which has not previously been possible. This methodology will facilitate re-analysis of archived microarray data and broaden the utility of the vast quantities of data still being generated.