Predicting base editing outcomes with an attention-based deep learning algorithm trained on high-throughput target library screens.
Ontology highlight
ABSTRACT:
SUBMITTER: Marquart KF
PROVIDER: S-EPMC8387386 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:In a recent issue of Cell, Martin-Rufino et al. develop a strategy for performing high-throughput base-editing CRISPR screens coupled with single-cell readouts in the context of human hematopoiesis. Through a series of proof-of-principle experiments, the authors demonstrate the potential of base-editing screens for the study and treatment of hematological disorders.
| S-EPMC10391558 | biostudies-literature
Project description:Mutations in DNA damage response (DDR) genes endanger genome integrity and predispose to cancer and genetic disorders. Here, using CRISPR-dependent cytosine base editing screens, we identify > 2,000 sgRNAs that generate nucleotide variants in 86 DDR genes, resulting in altered cellular fitness upon DNA damage. Among those variants, we discover loss- and gain-of-function mutants in the Tudor domain of the DDR regulator 53BP1 that define a non-canonical surface required for binding the deubiquitinase USP28. Moreover, we characterize variants of the TRAIP ubiquitin ligase that define a domain, whose loss renders cells resistant to topoisomerase I inhibition. Finally, we identify mutations in the ATM kinase with opposing genome stability phenotypes and loss-of-function mutations in the CHK2 kinase previously categorized as variants of uncertain significance for breast cancer. We anticipate that this resource will enable the discovery of additional DDR gene functions and expedite studies of DDR variants in human disease.
| S-EPMC8018281 | biostudies-literature
Project description:Interferon-γ (IFN-γ) signaling mediates host responses to infection, inflammation and anti-tumor immunity. Mutations in the IFN-γ signaling pathway cause immunological disorders, hematological malignancies, and resistance to immune checkpoint blockade (ICB) in cancer; however, the function of most clinically observed variants remains unknown. Here, we systematically investigate the genetic determinants of IFN-γ response in colorectal cancer cells using CRISPR-Cas9 screens and base editing mutagenesis. Deep mutagenesis of JAK1 with cytidine and adenine base editors, combined with pathway-wide screens, reveal loss-of-function and gain-of-function mutations, including causal variants in hematological malignancies and mutations detected in patients refractory to ICB. We functionally validate variants of uncertain significance in primary tumor organoids, where engineering missense mutations in JAK1 enhanced or reduced sensitivity to autologous tumor-reactive T cells. We identify more than 300 predicted missense mutations altering IFN-γ pathway activity, generating a valuable resource for interpreting gene variant function.
| S-EPMC9942875 | biostudies-literature
Project description:BackgroundMillions of nucleotide variants are identified through cancer genome sequencing and it is clinically important to identify the pathogenic variants among them. By introducing base substitutions at guide RNA target regions in the genome, CRISPR-Cas9-based base editors provide the possibility for evaluating a large number of variants in their genomic context. However, the variability in editing efficiency and the complexity of outcome mapping are two existing problems for assigning guide RNA effects to variants in base editing screens.ResultsTo improve the identification of pathogenic variants, we develop a framework to combine base editing screens with sgRNA efficiency and outcome mapping. We apply the method to evaluate more than 9000 variants across all the exons of BRCA1 and BRCA2 genes. Our efficiency-corrected scoring model identifies 910 loss-of-function variants for BRCA1/2, including 151 variants in the noncoding part of the genes such as the 5' untranslated regions. Many of them are identified in cancer patients and are reported as "benign/likely benign" or "variants of uncertain significance" by clinicians. Our data suggest a need to re-evaluate their clinical significance, which may be helpful for risk assessment and treatment of breast and ovarian cancer.ConclusionsOur results suggest that base editing screens with efficiency correction is a powerful strategy to identify pathogenic variants in a high-throughput manner. Applying this strategy to assess variants in both coding and noncoding regions of the genome could have a direct impact on the interpretation of cancer variants.
| S-EPMC7945310 | biostudies-literature
Project description:Recent advances in multiplexed single-cell transcriptomics experiments facilitate the high-throughput study of drug and genetic perturbations. However, an exhaustive exploration of the combinatorial perturbation space is experimentally unfeasible. Therefore, computational methods are needed to predict, interpret, and prioritize perturbations. Here, we present the compositional perturbation autoencoder (CPA), which combines the interpretability of linear models with the flexibility of deep-learning approaches for single-cell response modeling. CPA learns to in silico predict transcriptional perturbation response at the single-cell level for unseen dosages, cell types, time points, and species. Using newly generated single-cell drug combination data, we validate that CPA can predict unseen drug combinations while outperforming baseline models. Additionally, the architecture's modularity enables incorporating the chemical representation of the drugs, allowing the prediction of cellular response to completely unseen drugs. Furthermore, CPA is also applicable to genetic combinatorial screens. We demonstrate this by imputing in silico 5,329 missing combinations (97.6% of all possibilities) in a single-cell Perturb-seq experiment with diverse genetic interactions. We envision CPA will facilitate efficient experimental design and hypothesis generation by enabling in silico response prediction at the single-cell level and thus accelerate therapeutic applications using single-cell technologies.
| S-EPMC10258562 | biostudies-literature
Project description:CRISPR/Cas base editors promise nucleotide-level control over DNA sequences, but the determinants of their activity remain incompletely understood. We measured base editing frequencies in two human cell lines for two cytosine and two adenine base editors at ∼14 000 target sequences and find that base editing activity is sequence-biased, with largest effects from nucleotides flanking the target base. Whether a base is edited depends strongly on the combination of its position in the target and the preceding base, acting to widen or narrow the effective editing window. The impact of features on editing rate depends on the position, with sequence bias efficacy mainly influencing bases away from the center of the window. We use these observations to train a machine learning model to predict editing activity per position, with accuracy ranging from 0.49 to 0.72 between editors, and with better generalization across datasets than existing tools. We demonstrate the usefulness of our model by predicting the efficacy of disease mutation correcting guides, and find that most of them suffer from more unwanted editing than pure outcomes. This work unravels the position-specificity of base editing biases and allows more efficient planning of editing campaigns in experimental and therapeutic contexts.
| S-EPMC8989541 | biostudies-literature
Project description:Although base editors are widely used to install targeted point mutations, the factors that determine base editing outcomes are not well understood. We characterized sequence-activity relationships of 11 cytosine and adenine base editors (CBEs and ABEs) on 38,538 genomically integrated targets in mammalian cells and used the resulting outcomes to train BE-Hive, a machine learning model that accurately predicts base editing genotypic outcomes (R ≈ 0.9) and efficiency (R ≈ 0.7). We corrected 3,388 disease-associated SNVs with ≥90% precision, including 675 alleles with bystander nucleotides that BE-Hive correctly predicted would not be edited. We discovered determinants of previously unpredictable C-to-G, or C-to-A editing and used these discoveries to correct coding sequences of 174 pathogenic transversion SNVs with ≥90% precision. Finally, we used insights from BE-Hive to engineer novel CBE variants that modulate editing outcomes. These discoveries illuminate base editing, enable editing at previously intractable targets, and provide new base editors with improved editing capabilities.
| S-EPMC7384975 | biostudies-literature
Project description:The current-generation adenine base editor (ABE) ABE8e, which has evolved from the prokaryotic evolution system, exhibits high efficiency in mediating A-to-G conversion and is presumed to be promising for gene therapy. However, its much wider editing window and substantially higher off-target editing activity restricted its applications in precise base editing for therapeutic use. This study uses a library-assisted protein evolution approach using eukaryotic cells to generate ABE variants with improved specificity and reduced off-target editing while maintaining high activity in human cells. The study generated an expanded set of ABEs with efficient editing activities and chose four evolved variants that offered either similar or modestly higher efficiency within a narrower editing window of protospacer position ≈4-7 compared to that of ABE8e in human cells, which would enable minimized bystander editing. Moreover, these variants resulted in reduced off-target editing events when delivered as plasmid or mRNA into human cells. Finally, these variants can install both disease-suppressing mutations and disease-correcting mutations efficiently with minimal undesired bystander editing making them promising approaches for specific therapeutic edits. In summary, the work establishes a mutant-library-assisted protein evolution method in eukaryotic cells and generates alternative ABE variants as efficient tools for precise human genome editing.
| S-EPMC11321652 | biostudies-literature
Project description:Prime editing is a versatile genome editing tool but requires experimental optimization of the prime editing guide RNA (pegRNA) to achieve high editing efficiency. Here we conducted a high-throughput screen to analyze prime editing outcomes of 92,423 pegRNAs on a highly diverse set of 13,349 human pathogenic mutations that include base substitutions, insertions and deletions. Based on this dataset, we identified sequence context features that influence prime editing and trained PRIDICT (prime editing guide prediction), an attention-based bidirectional recurrent neural network. PRIDICT reliably predicts editing rates for all small-sized genetic changes with a Spearman's R of 0.85 and 0.78 for intended and unintended edits, respectively. We validated PRIDICT on endogenous editing sites as well as an external dataset and showed that pegRNAs with high (>70) versus low (<70) PRIDICT scores showed substantially increased prime editing efficiencies in different cell types in vitro (12-fold) and in hepatocytes in vivo (tenfold), highlighting the value of PRIDICT for basic and for translational research applications.
| S-EPMC7614945 | biostudies-literature
Project description:As a simple and programmable nuclease-based genome editing tool, the CRISPR/Cas9 system has been widely used in target-gene repair and gene-expression regulation. The DNA mutation generated by CRISPR/Cas9-mediated double-strand breaks determines its biological and phenotypic effects. Experiments have demonstrated that CRISPR/Cas9-generated cellular-repair outcomes depend on local sequence features. Therefore, the repair outcomes after DNA break can be predicted by sequences near the cleavage sites. However, existing prediction methods rely on manually constructed features or insufficiently detailed prediction labels. They cannot satisfy clinical-level-prediction accuracy, which limit the performance of these models to existing knowledge about CRISPR/Cas9 editing. We predict 557 repair labels of DNA, covering the vast majority of Cas9-generated mutational outcomes, and build a deep learning model called Apindel, to predict CRISPR/Cas9 editing outcomes. Apindel, automatically, trains the sequence features of DNA with the GloVe model, introduces location information through Positional Encoding (PE), and embeds the trained-word vector matrixes into a deep learning model, containing BiLSTM and the Attention mechanism. Apindel has better performance and more detailed prediction categories than the most advanced DNA-mutation-predicting models. It, also, reveals that nucleotides at different positions relative to the cleavage sites have different influences on CRISPR/Cas9 editing outcomes.
| S-EPMC9180579 | biostudies-literature