Predicting Splicing from Primary Sequence with Deep Learning.

The splicing of pre-mRNAs into mature transcripts is remarkable for its precision, but the mechanisms by which the cellular machinery achieves such specificity are incompletely understood. Here, we describe a deep neural network that accurately predicts splice junctions from an arbitrary pre-mRNA transcript sequence, enabling precise prediction of noncoding genetic variants that cause cryptic splicing. Synonymous and intronic mutations with predicted splice-altering consequence validate at a high rate on RNA-seq and are strongly deleterious in the human population. De novo mutations with predicted splice-altering consequence are significantly enriched in patients with autism and intellectual disability compared to healthy controls and validate against RNA-seq in 21 out of 28 of these patients. We estimate that 9%-11% of pathogenic mutations in patients with rare genetic disorders are caused by this previously underappreciated class of disease variation.

Fine-mapping inflammatory bowel disease loci to single-variant resolution.

Inflammatory bowel diseases are chronic gastrointestinal inflammatory disorders that affect millions of people worldwide. Genome-wide association studies have identified 200 inflammatory bowel disease-associated loci, but few have been conclusively resolved to specific functional variants. Here we report fine-mapping of 94 inflammatory bowel disease loci using high-density genotyping in 67,852 individuals. We pinpoint 18 associations to a single causal variant with greater than 95% certainty, and an additional 27 associations to a single variant with greater than 50% certainty. These 45 variants are significantly enriched for protein-coding changes (n = 13), direct disruption of transcription-factor binding sites (n = 3), and tissue-specific epigenetic marks (n = 10), with the last category showing enrichment in specific immune cells among associations stronger in Crohn's disease and in gut mucosa among associations stronger in ulcerative colitis. The results of this study suggest that high-resolution fine-mapping in large samples can convert many discoveries from genome-wide association studies into statistically convincing causal variants, providing a powerful substrate for experimental elucidation of disease mechanisms.

PICS2: next-generation fine mapping via probabilistic identification of causal SNPs.

SUMMARY: The Probabilistic Identification of Causal SNPs (PICS) algorithm and web application was developed as a fine-mapping tool to determine the likelihood that each single nucleotide polymorphism (SNP) in LD with a reported index SNP is a true causal polymorphism. PICS is notable for its ability to identify candidate causal SNPs within a locus using only the index SNP, which are widely available from published GWAS, whereas other methods require full summary statistics or full genotype data. However, the original PICS web application operates on a single SNP at a time, with slow performance, severely limiting its usability. We have developed a next-generation PICS tool, PICS2, which enables performance of PICS analyses of large batches of index SNPs with much faster performance. Additional updates and extensions include use of LD reference data generated from 1000 Genomes phase 3; annotation of variant consequences; annotation of GTEx eQTL genes and downloadable PICS SNPs from GTEx eQTLs; the option of generating PICS probabilities from experimental summary statistics; and generation of PICS SNPs from all SNPs of the GWAS catalog, automatically updated weekly. These free and easy-to-use resources will enable efficient determination of candidate loci for biological studies to investigate the true causal variants underlying disease processes. AVAILABILITY AND IMPLEMENTATION: PICS2 is available at https://pics2.ucsf.edu. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Genetic and epigenetic fine mapping of causal autoimmune disease variants.

Genome-wide association studies have identified loci underlying human diseases, but the causal nucleotide changes and mechanisms remain largely unknown. Here we developed a fine-mapping algorithm to identify candidate causal variants for 21 autoimmune diseases from genotyping data. We integrated these predictions with transcription and cis-regulatory element annotations, derived by mapping RNA and chromatin in primary immune cells, including resting and stimulated CD4(+) T-cell subsets, regulatory T cells, CD8(+) T cells, B cells, and monocytes. We find that ∼90% of causal variants are non-coding, with ∼60% mapping to immune-cell enhancers, many of which gain histone acetylation and transcribe enhancer-associated RNA upon immune stimulation. Causal variants tend to occur near binding sites for master regulators of immune differentiation and stimulus-dependent gene activation, but only 10-20% directly alter recognizable transcription factor binding motifs. Rather, most non-coding risk variants, including those that alter gene expression, affect non-canonical sequence determinants not well-explained by current gene regulatory models.

Variation in predicted COVID-19 risk among lemurs and lorises.

The novel coronavirus SARS-CoV-2, which in humans leads to the disease COVID-19, has caused global disruption and more than 2 million fatalities since it first emerged in late 2019. As we write, infection rates are at their highest point globally and are rising extremely rapidly in some areas due to more infectious variants. The primary target of SARS-CoV-2 is the cellular receptor angiotensin-converting enzyme-2 (ACE2). Recent sequence analyses of the ACE2 gene predict that many nonhuman primates are also likely to be highly susceptible to infection. However, the anticipated risk is not equal across the Order. Furthermore, some taxonomic groups show high ACE2 amino acid conservation, while others exhibit high variability at this locus. As an example of the latter, analyses of strepsirrhine primate ACE2 sequences to date indicate large variation among lemurs and lorises compared to other primate clades despite low sampling effort. Here, we report ACE2 gene and protein sequences for 71 individual strepsirrhines, spanning 51 species and 19 genera. Our study reinforces previous results while finding additional variability in other strepsirrhine species, and suggests several clades of lemurs have high potential susceptibility to SARS-CoV-2 infection. Troublingly, some species, including the rare and endangered aye-aye (Daubentonia madagascariensis), as well as those in the genera Avahi and Propithecus, may be at high risk. Given that lemurs are endemic to Madagascar and among the primates at highest risk of extinction globally, further understanding of the potential threat of COVID-19 to their health should be a conservation priority. All feasible actions should be taken to limit their exposure to SARS-CoV-2.

Expanding the microRNA targeting code: functional sites with centered pairing.

Most metazoan microRNA (miRNA) target sites have perfect pairing to the seed region, located near the miRNA 5' end. Although pairing to the 3' region sometimes supplements seed matches or compensates for mismatches, pairing to the central region has been known to function only at rare sites that impart Argonaute-catalyzed mRNA cleavage. Here, we present "centered sites," a class of miRNA target sites that lack both perfect seed pairing and 3'-compensatory pairing and instead have 11-12 contiguous Watson-Crick pairs to the center of the miRNA. Although centered sites can impart mRNA cleavage in vitro (in elevated Mg(2+)), in cells they repress protein output without consequential Argonaute-catalyzed cleavage. Our study also identified extensively paired sites that are cleavage substrates in cultured cells and human brain. This expanded repertoire of cleavage targets and the identification of the centered site type help explain why central regions of many miRNAs are evolutionarily conserved.

Whole genomes of Amazonian uakari monkeys reveal complex connectivity and fast differentiation driven by high environmental dynamism.

Despite showing the greatest primate diversity on the planet, genomic studies on Amazonian primates show very little representation in the literature. With 48 geolocalized high coverage whole genomes from wild uakari monkeys, we present the first population-level study on platyrrhines using whole genome data. In a very restricted range of the Amazon rainforest, eight uakari species (Cacajao genus) have been described and categorized into the bald and black uakari groups, based on phenotypic and ecological differences. Despite a slight habitat overlap, we show that posterior to their split 0.92 Mya, bald and black uakaris have remained independent, without gene flow. Nowadays, these two groups present distinct genetic diversity and group-specific variation linked to pathogens. We propose differing hydrology patterns and effectiveness of geographic barriers have modulated the intra-group connectivity and structure of bald and black uakari populations. With this work we have explored the effects of the Amazon rainforest's dynamism on wild primates' genetics and increased the representation of platyrrhine genomes, thus opening the door to future research on the complexity and diversity of primate genomics.

Single-cell chromatin accessibility reveals malignant regulatory programs in primary human cancers.

To identify cancer-associated gene regulatory changes, we generated single-cell chromatin accessibility landscapes across eight tumor types as part of The Cancer Genome Atlas. Tumor chromatin accessibility is strongly influenced by copy number alterations that can be used to identify subclones, yet underlying cis-regulatory landscapes retain cancer type-specific features. Using organ-matched healthy tissues, we identified the "nearest healthy" cell types in diverse cancers, demonstrating that the chromatin signature of basal-like-subtype breast cancer is most similar to secretory-type luminal epithelial cells. Neural network models trained to learn regulatory programs in cancer revealed enrichment of model-prioritized somatic noncoding mutations near cancer-associated genes, suggesting that dispersed, nonrecurrent, noncoding mutations in cancer are functional. Overall, these data and interpretable gene regulatory models for cancer and healthy tissue provide a framework for understanding cancer-specific gene regulation.

Rare penetrant mutations confer severe risk of common diseases.

We examined 454,712 exomes for genes associated with a wide spectrum of complex traits and common diseases and observed that rare, penetrant mutations in genes implicated by genome-wide association studies confer ~10-fold larger effects than common variants in the same genes. Consequently, an individual at the phenotypic extreme and at the greatest risk for severe, early-onset disease is better identified by a few rare penetrant variants than by the collective action of many common variants with weak effects. By combining rare variants across phenotype-associated genes into a unified genetic risk model, we demonstrate superior portability across diverse global populations compared with common-variant polygenic risk scores, greatly improving the clinical utility of genetic-based risk prediction.

Phylogenomic analyses provide insights into primate evolution.

Comparative analysis of primate genomes within a phylogenetic context is essential for understanding the evolution of human genetic architecture and primate diversity. We present such a study of 50 primate species spanning 38 genera and 14 families, including 27 genomes first reported here, with many from previously less well represented groups, the New World monkeys and the Strepsirrhini. Our analyses reveal heterogeneous rates of genomic rearrangement and gene evolution across primate lineages. Thousands of genes under positive selection in different lineages play roles in the nervous, skeletal, and digestive systems and may have contributed to primate innovations and adaptations. Our study reveals that many key genomic innovations occurred in the Simiiformes ancestral node and may have had an impact on the adaptive radiation of the Simiiformes and human evolution.

Variation in predicted COVID-19 risk among lemurs and lorises.

The novel coronavirus SARS-CoV-2, which in humans leads to the disease COVID-19, has caused global disruption and more than 1.5 million fatalities since it first emerged in late 2019. As we write, infection rates are currently at their highest point globally and are rising extremely rapidly in some areas due to more infectious variants. The primary viral target is the cellular receptor angiotensin-converting enzyme-2 (ACE2). Recent sequence analyses of the ACE2 gene predicts that many nonhuman primates are also likely to be highly susceptible to infection. However, the anticipated risk is not equal across the Order. Furthermore, some taxonomic groups show high ACE2 amino acid conservation, while others exhibit high variability at this locus. As an example of the latter, analyses of strepsirrhine primate ACE2 sequences to date indicate large variation among lemurs and lorises compared to other primate clades despite low sampling effort. Here, we report ACE2 gene and protein sequences for 71 individual strepsirrhines, spanning 51 species and 19 genera. Our study reinforces previous results and finds additional variability in other strepsirrhine species, and suggests several clades of lemurs have high potential susceptibility to SARS-CoV-2 infection. Troublingly, some species, including the rare and Endangered aye-aye (Daubentonia madagascariensis), as well as those in the genera Avahi and Propithecus, may be at high risk. Given that lemurs are endemic to Madagascar and among the primates at highest risk of extinction globally, further understanding of the potential threat of COVID-19 to their health should be a conservation priority. All feasible actions should be taken to limit their exposure to SARS-CoV-2.

Clinical impact of splicing in neurodevelopmental disorders.

Clinical exome sequencing is frequently used to identify gene-disrupting variants in individuals with neurodevelopmental disorders. While splice-disrupting variants are known to contribute to these disorders, clinical interpretation of cryptic splice variants outside of the canonical splice site has been challenging. Here, we discuss papers that improve such detection.

Author Correction: Predicting the clinical impact of human mutation with deep neural networks.

In the version of this article originally published, the name of author Serafim Batzoglou was misspelled. The error has been corrected in the HTML and PDF versions of the article.

Predicting the clinical impact of human mutation with deep neural networks.

Millions of human genomes and exomes have been sequenced, but their clinical applications remain limited due to the difficulty of distinguishing disease-causing mutations from benign genetic variation. Here we demonstrate that common missense variants in other primate species are largely clinically benign in human, enabling pathogenic mutations to be systematically identified by the process of elimination. Using hundreds of thousands of common variants from population sequencing of six non-human primate species, we train a deep neural network that identifies pathogenic mutations in rare disease patients with 88% accuracy and enables the discovery of 14 new candidate genes in intellectual disability at genome-wide significance. Cataloging common variation from additional primate species would improve interpretation for millions of variants of uncertain significance, further advancing the clinical utility of human genome sequencing.

Genome-wide identification and characterization of functional neuronal activity-dependent enhancers.

Experience-dependent gene transcription is required for nervous system development and function. However, the DNA regulatory elements that control this program of gene expression are not well defined. Here we characterize the enhancers that function across the genome to mediate activity-dependent transcription in mouse cortical neurons. We find that the subset of enhancers enriched for monomethylation of histone H3 Lys4 (H3K4me1) and binding of the transcriptional coactivator CREBBP (also called CBP) that shows increased acetylation of histone H3 Lys27 (H3K27ac) after membrane depolarization of cortical neurons functions to regulate activity-dependent transcription. A subset of these enhancers appears to require binding of FOS, which was previously thought to bind primarily to promoters. These findings suggest that FOS functions at enhancers to control activity-dependent gene programs that are critical for nervous system function and provide a resource of functional cis-regulatory elements that may give insight into the genetic variants that contribute to brain development and disease.

Most mammalian mRNAs are conserved targets of microRNAs.

MicroRNAs (miRNAs) are small endogenous RNAs that pair to sites in mRNAs to direct post-transcriptional repression. Many sites that match the miRNA seed (nucleotides 2-7), particularly those in 3' untranslated regions (3'UTRs), are preferentially conserved. Here, we overhauled our tool for finding preferential conservation of sequence motifs and applied it to the analysis of human 3'UTRs, increasing by nearly threefold the detected number of preferentially conserved miRNA target sites. The new tool more efficiently incorporates new genomes and more completely controls for background conservation by accounting for mutational biases, dinucleotide conservation rates, and the conservation rates of individual UTRs. The improved background model enabled preferential conservation of a new site type, the "offset 6mer," to be detected. In total, >45,000 miRNA target sites within human 3'UTRs are conserved above background levels, and >60% of human protein-coding genes have been under selective pressure to maintain pairing to miRNAs. Mammalian-specific miRNAs have far fewer conserved targets than do the more broadly conserved miRNAs, even when considering only more recently emerged targets. Although pairing to the 3' end of miRNAs can compensate for seed mismatches, this class of sites constitutes less than 2% of all preferentially conserved sites detected. The new tool enables statistically powerful analysis of individual miRNA target sites, with the probability of preferentially conserved targeting (P(CT)) correlating with experimental measurements of repression. Our expanded set of target predictions (including conserved 3'-compensatory sites), are available at the TargetScan website, which displays the P(CT) for each site and each predicted target.

MicroRNA targeting specificity in mammals: determinants beyond seed pairing.

Mammalian microRNAs (miRNAs) pair to 3'UTRs of mRNAs to direct their posttranscriptional repression. Important for target recognition are approximately 7 nt sites that match the seed region of the miRNA. However, these seed matches are not always sufficient for repression, indicating that other characteristics help specify targeting. By combining computational and experimental approaches, we uncovered five general features of site context that boost site efficacy: AU-rich nucleotide composition near the site, proximity to sites for coexpressed miRNAs (which leads to cooperative action), proximity to residues pairing to miRNA nucleotides 13-16, positioning within the 3'UTR at least 15 nt from the stop codon, and positioning away from the center of long UTRs. A model combining these context determinants quantitatively predicts site performance both for exogenously added miRNAs and for endogenous miRNA-message interactions. Because it predicts site efficacy without recourse to evolutionary conservation, the model also identifies effective nonconserved sites and siRNA off-targets.

The widespread impact of mammalian MicroRNAs on mRNA repression and evolution.

Thousands of mammalian messenger RNAs are under selective pressure to maintain 7-nucleotide sites matching microRNAs (miRNAs). We found that these conserved targets are often highly expressed at developmental stages before miRNA expression and that their levels tend to fall as the miRNA that targets them begins to accumulate. Nonconserved sites, which outnumber the conserved sites 10 to 1, also mediate repression. As a consequence, genes preferentially expressed at the same time and place as a miRNA have evolved to selectively avoid sites matching the miRNA. This phenomenon of selective avoidance extends to thousands of genes and enables spatial and temporal specificities of miRNAs to be revealed by finding tissues and developmental stages in which messages with corresponding sites are expressed at lower levels.