Massively parallel identification of zipcodes in primary cortical neurons

Cells adopt highly polarized shapes and form distinct subcellular compartments largely due to the localization of many mRNAs to specific areas, where they are translated into proteins with local functions. This mRNA localization is mediated by specific cis-regulatory elements in mRNAs, commonly called "zipcodes." Their recognition by RNA-binding proteins (RBPs) leads to the integration of the mRNAs into macromolecular complexes and their localization. While there are hundreds of localized mRNAs, only a few zipcodes have been characterized. Here, we describe a novel neuronal zipcode identification protocol (N-zip) that can identify zipcodes across hundreds of 3'UTRs. This approach combines a method of separating the principal subcellular compartments of neurons – cell bodies and neurites - with a massively parallel reporter assay. Our analysis identifies the let-7 binding site and (AU)n motif as de novo zipcodes in mouse primary cortical neurons and suggests a strategy for detecting many more.

Massively parallel analysis of human 3′ UTRs reveals that AU-rich element length and registration predict mRNA destabilization

AU-rich elements (AREs) are 3′ UTR cis-regulatory elements that regulate the stability of mRNAs. Consensus ARE motifs have been determined, but little is known about how differences in 3′ UTR sequences that conform to these motifs affect their function. Here we use functional annotation of sequences from 3′ UTRs (fast-UTR), a massively parallel reporter assay (MPRA), to investigate the effects of 41,288 3′ UTR sequence fragments from 4,653 transcripts on gene expression and mRNA stability in Jurkat and Beas2B cells. Our analyses demonstrate that the length of an ARE and its registration (the first and last nucleotides of the repeating ARE motif) have significant effects on gene expression and stability. Based on this finding, we propose improved ARE classification and concomitant methods to categorize and predict the effect of AREs on gene expression and stability. Finally, to investigate the advantages of our general experimental design we examine other motifs including constitutive decay elements (CDEs), where we show that the length of the CDE stem-loop has a significant impact on steady-state expression and mRNA stability. We conclude that fast-UTR, in conjunction with our analytical approach, can produce improved yet simple sequence-based rules for predicting the activity of human 3′ UTRs.

A single-cell massively parallel reporter assay detects cell type specific cis-regulatory activity

We developed a single-cell massively parallel reporter assay (scMPRA) to measure the activity of libraries of cis-regulatory sequences (CRSs) across multiple cell-types simultaneously. As a proof of concept, we assayed a library of core promoters in a mixture of HEK293 and K562 cells and showed that scMPRA is a reproducible, highly parallel, single-cell reporter gene assay. Our results show that housekeeping promoters and CpG island promoters have lower activity in K562 cells relative to HEK293, which likely reflects developmental differences between the cell lines. Within K562 cells, scMPRA identified a subset of developmental promoters that are upregulated in the CD34+ /CD38 - sub-state, confirming this state as more "stem-like." Finally, we deconvolved the intrinsic and extrinsic components of promoter cell-to-cell variability and found that developmental promoters have a higher proportion of extrinsic noise compared to housekeeping promoters, which may reflect the responsiveness of developmental promoters to the cellular environment. We anticipate scMPRA will be widely applicable for studying the role of CRSs across diverse cell types.

Massively parallel identification of functionally consequential noncoding genetic variants in undiagnosed rare disease patients

Clinical whole genome sequencing has enabled the discovery of potentially pathogenic noncoding variants in the genomes of rare disease patients with a prior history of negative genetic testing. However, interpreting the functional consequences of noncoding variants and distinguishing those that contribute to disease etiology remains a challenge. Here we address this challenge by experimentally profiling the functional consequences of rare noncoding variants detected in a cohort of undiagnosed rare disease patients at scale using a massively parallel reporter assay. We demonstrate that this approach successfully identifies rare noncoding variants that alter the regulatory capacity of genomic sequences. In addition, we describe an integrative analysis that utilizes genomic features alongside patient clinical data to further prioritize candidate variants with an increased likelihood of pathogenicity. This work represents an important step towards establishing a framework for the functional interpretation of clinically detected noncoding variants.

Massively parallel identification of zipcodes in primary cortical neurons

Cells adopt highly polarized shapes and form distinct subcellular compartments largely due to the localization of many mRNAs to specific areas, where they are translated into proteins with local functions. This mRNA localization is mediated by specific cis-regulatory elements in mRNAs, commonly called "zipcodes." Their recognition by RNA-binding proteins (RBPs) leads to the integration of the mRNAs into macromolecular complexes and their localization. While there are hundreds of localized mRNAs, only a few zipcodes have been characterized. Here, we describe a novel neuronal zipcode identification protocol (N-zip) that can identify zipcodes across hundreds of 3'UTRs. This approach combines a method of separating the principal subcellular compartments of neurons - cell bodies and neurites - with a massively parallel reporter assay. Our analysis identifies the let-7 binding site and (AU)n motif as de novo zipcodes in mouse primary cortical neurons and suggests a strategy for detecting many more.

A functional genomics pipeline identifies pleiotropy and cross-tissue effects within obesity-associated GWAS loci

Genome-wide association studies (GWAS) have identified many disease-associated variants, yet mechanisms underlying these associations remain unclear. To understand obesity-associated variants, we generate gene regulatory annotations in adipocytes and hypothalamic neurons across cellular differentiation stages. We then test variants in 97 obesity-associated loci using a massively parallel reporter assay and identify putatively causal variants that display cell type specific or cross-tissue enhancer-modulating properties. Integrating these variants with gene regulatory information suggests genes that underlie obesity GWAS associations. We also investigate a complex genomic interval on 16p11.2 where two independent loci exhibit megabase-range, cross-locus chromatin interactions. We demonstrate that variants within these two loci regulate a shared gene set. Together, our data support a model where GWAS loci contain variants that alter enhancer activity across tissues, potentially with temporally restricted effects, to impact the expression of multiple genes. This complex model has broad implications for ongoing efforts to understand GWAS.

Converging evidence for differential regulatory control of APOEε4 on African versus European haplotypes

INTRODUCTION. The difference in APOEϵ4 risk for Alzheimer disease (AD) between different populations is associated with APOEϵ4 local ancestry (LA). We examined LA SNPs with significant frequency differences between African and European/Japanese APOEϵ4 haplotypes for areas of differential regulation. METHODS. We performed two enhancer Massively Parallel Reporter Assay (MPRA) approaches, supplemented with single fragment reporter assays. We utilized Capture C analyses to support interactions with the APOE promoter. RESULTS. The TOMM40 intron 2 and 3 region showed increased enhancer activity in the European/Japanese versus African LA haplotypes in astrocytes and microglia. This region overlaps with APOE promoter interactions as assessed by Capture C analysis. Single variant analyses pinpoints rs2075650/rs157581, and rs59007384 as functionally different on these haplotypes. DISCUSSION. Both differential regulatory function and Capture C data support an intronic region in TOMM40 as contributing to the differential APOE expression between African and European/Japanese LA.

Analysis of long and short enhancers in melanoma cell states

Understanding how enhancers drive cell type specificity and efficiently identifying them is essential for the development of innovative therapeutic strategies. In melanoma, the melanocytic (MEL) and the mesenchymal-like (MES) states present themselves with different responses to therapy, making the identification of specific enhancers highly relevant. Using massively parallel reporter assay (MPRA) in a panel of patient-derived melanoma lines (MM lines), we set to identify and decipher melanoma enhancers by first focusing on regions with state specific H3K27 acetylation close to differentially expressed genes. A more in-depth evaluation of those regions was then pursued by investigating the activity of ATAC-seq peaks found therein along with a full tiling of the acetylated regions with 190 bp sequences. Activity was observed in more than 60% of the selected regions and we were able to precisely locate the active regions within ATAC-seq peaks. Comparison of sequence content with activity, using the deep learning model DeepMEL2, revealed that AP-1 alone is responsible for the MES enhancer activity, while SOX and MITF both influence MEL enhancer activity with SOX being required to achieve high levels of activity. Overall, our MPRA assays shed light on the relationship between long and short sequences in terms of their sequence content, enhancer activity, and specificity as reporters across melanoma cell states.

Regulatory dissection of the severe COVID-19 risk locus introgressed by Neanderthals

Individuals infected with the SARS-CoV-2 virus present with a wide variety of phenotypes ranging from asymptomatic to severe and even lethal outcomes. Past research has revealed a genetic haplotype on chromosome 3 that entered the human population via introgression from Neanderthals as the strongest genetic risk factor for the severe COVID-19 phenotype. However, the specific variants along this introgressed haplotype that contribute to this risk and the biological mechanisms that are involved remain unclear. Here, we assess the variants present on the risk haplotype for their likelihood of driving the severe COVID-19 phenotype. We do this by first exploring their impact on the regulation of genes involved in COVID-19 infection using a variety of population genetics and functional genomics tools. We then perform an locus-specific massively parallel reporter assay to individually assess the regulatory potential of each allele on the haplotype in a multipotent immune-related cell line. We ultimately reduce the set of over 600 linked genetic variants to identify 4 introgressed alleles that are strong functional candidates for driving the association between this locus and severe COVID-19. These variants likely drive the locus impact on severity by putatively modulating the regulation of two critical chemokine receptor genes: CCR1 and CCR5. These alleles are ideal targets for future functional investigations into the interaction between host genomics and COVID-19 outcomes.

Predicting mean ribosome load for 5’UTR of any length using deep learning

The 5’ untranslated region plays a key role in regulating mRNA translation and consequently protein abundance. Therefore, accurate modeling of 5’UTR regulatory sequences shall provide insights into translational control mechanisms and help interpret genetic variants. Recently, a model was trained on a massively parallel reporter assay to predict mean ribosome load (MRL)—a proxy for translation rate—directly from 5’UTR sequence with a high degree of accuracy. However, this model is restricted to sequence lengths investigated in the reporter assay and therefore cannot be applied to the majority of human sequences without a substantial loss of information. Here, we introduced frame pooling, a novel neural network operation that enabled the development of an MRL prediction model for 5’UTRs of any length. Our model shows state-of-the-art performance on fixed length randomized sequences, while offering better generalization performance on longer sequences and on a variety of translation-related genome-wide datasets. Variant interpretation is demonstrated on a 5’UTR variant of the gene HBB associated with beta-thalassemia. Frame pooling could find applications in other bioinformatics predictive tasks. Moreover, our model, released open source, could help pinpoint pathogenic genetic variants.