Anti-CRISPR prediction using deep learning reveals an inhibitor of Cas13b nucleases.

As part of the ongoing bacterial-phage arms race, CRISPR-Cas systems in bacteria clear invading phages whereas anti-CRISPR proteins (Acrs) in phages inhibit CRISPR defenses. Known Acrs have proven extremely diverse, complicating their identification. Here, we report a deep learning algorithm for Acr identification that revealed an Acr against type VI-B CRISPR-Cas systems. The algorithm predicted numerous putative Acrs spanning almost all CRISPR-Cas types and subtypes, including over 7,000 putative type IV and VI Acrs not predicted by other algorithms. By performing a cell-free screen for Acr hits against type VI-B systems, we identified a potent inhibitor of Cas13b nucleases we named AcrVIB1. AcrVIB1 blocks Cas13b-mediated defense against a targeted plasmid and lytic phage, and its inhibitory function principally occurs upstream of ribonucleoprotein complex formation. Overall, our work helps expand the known Acr universe, aiding our understanding of the bacteria-phage arms race and the use of Acrs to control CRISPR technologies.

SpacerPlacer: ancestral reconstruction of CRISPR arrays reveals the evolutionary dynamics of spacer deletions.

Bacteria employ CRISPR-Cas systems for defense by integrating invader-derived sequences, termed spacers, into the CRISPR array, which constitutes an immunity memory. While spacer deletions occur randomly across the array, newly acquired spacers are predominantly integrated at the leader end. Consequently, spacer arrays can be used to derive the chronology of spacer insertions. Reconstruction of ancestral spacer acquisitions and deletions could help unravel the coevolution of phages and bacteria, the evolutionary dynamics in microbiomes, or track pathogens. However, standard reconstruction methods produce misleading results by overlooking insertion order and joint deletions of spacers. Here, we present SpacerPlacer, a maximum likelihood-based ancestral reconstruction approach for CRISPR array evolution. We used SpacerPlacer to reconstruct and investigate ancestral deletion events of 4565 CRISPR arrays, revealing that spacer deletions occur 374 times more frequently than mutations and are regularly deleted jointly, with an average of 2.7 spacers. Surprisingly, we observed a decrease in the spacer deletion frequency towards both ends of the reconstructed arrays. While the resulting trailer-end conservation is commonly observed, a reduced deletion frequency is now also detectable towards the variable leader end. Finally, our results point to the hypothesis that frequent loss of recently acquired spacers may provide a selective advantage.

Improved discovery of RNA-binding protein binding sites in eCLIP data using DEWSeq.

Enhanced crosslinking and immunoprecipitation (eCLIP) sequencing is a method for transcriptome-wide detection of binding sites of RNA-binding proteins (RBPs). However, identified crosslink sites can deviate from experimentally established functional elements of even well-studied RBPs. Current peak-calling strategies result in low replication and high false positive rates. Here, we present the R/Bioconductor package DEWSeq that makes use of replicate information and size-matched input controls. We benchmarked DEWSeq on 107 RBPs for which both eCLIP data and RNA sequence motifs are available and were able to more than double the number of motif-containing binding regions relative to standard eCLIP processing. The improvement not only relates to the number of binding sites (3.1-fold with known motifs for RBFOX2), but also their subcellular localization (1.9-fold of mitochondrial genes for FASTKD2) and structural targets (2.2-fold increase of stem-loop regions for SLBP. On several orthogonal CLIP-seq datasets, DEWSeq recovers a larger number of motif-containing binding sites (3.3-fold). DEWSeq is a well-documented R/Bioconductor package, scalable to adequate numbers of replicates, and tends to substantially increase the proportion and total number of RBP binding sites containing biologically relevant features.

Interrogating two extensively self-targeting Type I CRISPR-Cas systems in Xanthomonas albilineans reveals distinct anti-CRISPR proteins that block DNA degradation.

CRISPR-Cas systems store fragments of invader DNA as spacers to recognize and clear those same invaders in the future. Spacers can also be acquired from the host's genomic DNA, leading to lethal self-targeting. While self-targeting can be circumvented through different mechanisms, natural examples remain poorly explored. Here, we investigate extensive self-targeting by two CRISPR-Cas systems encoding 24 self-targeting spacers in the plant pathogen Xanthomonas albilineans. We show that the native I-C and I-F1 systems are actively expressed and that CRISPR RNAs are properly processed. When expressed in Escherichia coli, each Cascade complex binds its PAM-flanked DNA target to block transcription, while the addition of Cas3 paired with genome targeting induces cell killing. While exploring how X. albilineans survives self-targeting, we predicted putative anti-CRISPR proteins (Acrs) encoded within the bacterium's genome. Screening of identified candidates with cell-free transcription-translation systems and in E. coli revealed two Acrs, which we named AcrIC11 and AcrIF12Xal, that inhibit the activity of Cas3 but not Cascade of the respective system. While AcrF12Xal is homologous to AcrIF12, AcrIC11 shares sequence and structural homology with the anti-restriction protein KlcA. These findings help explain tolerance of self-targeting through two CRISPR-Cas systems and expand the known suite of DNA degradation-inhibiting Acrs.

Partial RNA design.

MOTIVATION: RNA design is a key technique to achieve new functionality in fields like synthetic biology or biotechnology. Computational tools could help to find such RNA sequences but they are often limited in their formulation of the search space. RESULTS: In this work, we propose partial RNA design, a novel RNA design paradigm that addresses the limitations of current RNA design formulations. Partial RNA design describes the problem of designing RNAs from arbitrary RNA sequences and structure motifs with multiple design goals. By separating the design space from the objectives, our formulation enables the design of RNAs with variable lengths and desired properties, while still allowing precise control over sequence and structure constraints at individual positions. Based on this formulation, we introduce a new algorithm, libLEARNA, capable of efficiently solving different constraint RNA design tasks. A comprehensive analysis of various problems, including a realistic riboswitch design task, reveals the outstanding performance of libLEARNA and its robustness. AVAILABILITY AND IMPLEMENTATION: libLEARNA is open-source and publicly available at: https://github.com/automl/learna_tools.

A mutually exclusive stem-loop arrangement in roX2 RNA is essential for X-chromosome regulation in Drosophila.

The X chromosome provides an ideal model system to study the contribution of RNA-protein interactions in epigenetic regulation. In male flies, roX long noncoding RNAs (lncRNAs) harbor several redundant domains to interact with the ubiquitin ligase male-specific lethal 2 (MSL2) and the RNA helicase Maleless (MLE) for X-chromosomal regulation. However, how these interactions provide the mechanics of spreading remains unknown. By using the uvCLAP (UV cross-linking and affinity purification) methodology, which provides unprecedented information about RNA secondary structures in vivo, we identified the minimal functional unit of roX2 RNA. By using wild-type and various MLE mutant derivatives, including a catalytically inactive MLE derivative, MLEGET, we show that the minimal roX RNA contains two mutually exclusive stem-loops that exist in a peculiar structural arrangement: When one stem-loop is unwound by MLE, an alternate structure can form, likely trapping MLE in this perpetually structured region. We show that this functional unit is necessary for dosage compensation, as mutations that disrupt this formation lead to male lethality. Thus, we propose that roX2 lncRNA contains an MLE-dependent affinity switch to enable reversible interactions of the MSL complex to allow dosage compensation of the X chromosome.

Heterogenous nuclear ribonucleoprotein D-like controls endothelial cell functions.

HnRNPs are ubiquitously expressed RNA-binding proteins, tightly controlling posttranscriptional gene regulation. Consequently, hnRNP networks are essential for cellular homeostasis and their dysregulation is associated with cancer and other diseases. However, the physiological function of hnRNPs in non-cancerous cell systems are poorly understood. We analyzed the importance of HNRNPDL in endothelial cell functions. Knockdown of HNRNPDL led to impaired proliferation, migration and sprouting of spheroids. Transcriptome analysis identified cyclin D1 (CCND1) and tropomyosin 4 (TPM4) as targets of HNRNPDL, reflecting the phenotypic changes after knockdown. Our findings underline the importance of HNRNPDL for the homeostasis of physiological processes in endothelial cells.

RiboReport - benchmarking tools for ribosome profiling-based identification of open reading frames in bacteria.

Small proteins encoded by short open reading frames (ORFs) with 50 codons or fewer are emerging as an important class of cellular macromolecules in diverse organisms. However, they often evade detection by proteomics or in silico methods. Ribosome profiling (Ribo-seq) has revealed widespread translation in genomic regions previously thought to be non-coding, driving the development of ORF detection tools using Ribo-seq data. However, only a handful of tools have been designed for bacteria, and these have not yet been systematically compared. Here, we aimed to identify tools that use Ribo-seq data to correctly determine the translational status of annotated bacterial ORFs and also discover novel translated regions with high sensitivity. To this end, we generated a large set of annotated ORFs from four diverse bacterial organisms, manually labeled for their translation status based on Ribo-seq data, which are available for future benchmarking studies. This set was used to investigate the predictive performance of seven Ribo-seq-based ORF detection tools (REPARATION_blast, DeepRibo, Ribo-TISH, PRICE, smORFer, ribotricer and SPECtre), as well as IRSOM, which uses coding potential and RNA-seq coverage only. DeepRibo and REPARATION_blast robustly predicted translated ORFs, including sORFs, with no significant difference for ORFs in close proximity to other genes versus stand-alone genes. However, no tool predicted a set of novel, experimentally verified sORFs with high sensitivity. Start codon predictions with smORFer show the value of initiation site profiling data to further improve the sensitivity of ORF prediction tools in bacteria. Overall, we find that bacterial tools perform well for sORF detection, although there is potential for improving their performance, applicability, usability and reproducibility.

RNA interaction format: a general data format for RNA interactions.

SUMMARY: RNA molecules play crucial roles in various biological processes. They mediate their function mainly by interacting with other RNAs or proteins. At present, information about these interactions is distributed over different resources, often providing the data in simple tab-delimited formats that differ between the databases. There is no standardized data format that can capture the nature of all these different interactions in detail. AVAILABILITY AND IMPLEMENTATION: Here, we propose the RNA interaction format (RIF) for the detailed representation of RNA-RNA and RNA-Protein interactions and provide reference implementations in C/C++, Python, and JavaScript. RIF is released under licence GNU General Public License version 3 (GNU GPLv3) and is available on https://github.com/RNABioInfo/rna-interaction-format.

Genome wide CRISPR screen for Pasteurella multocida toxin (PMT) binding proteins reveals LDL Receptor Related Protein 1 (LRP1) as crucial cellular receptor.

PMT is a protein toxin produced by Pasteurella multocida serotypes A and D. As causative agent of atrophic rhinitis in swine, it leads to rapid degradation of the nasal turbinate bone. The toxin acts as a deamidase to modify a crucial glutamine in heterotrimeric G proteins, which results in constitutive activation of the G proteins and permanent stimulation of numerous downstream signaling pathways. Using a lentiviral based genome wide CRISPR knockout screen in combination with a lethal toxin chimera, consisting of full length inactive PMT and the catalytic domain of diphtheria toxin, we identified the LRP1 gene encoding the Low-Density Lipoprotein Receptor-related protein 1 as a critical host factor for PMT function. Loss of LRP1 reduced PMT binding and abolished the cellular response and deamidation of heterotrimeric G proteins, confirming LRP1 to be crucial for PMT uptake. Expression of LRP1 or cluster 4 of LRP1 restored intoxication of the knockout cells. In summary our data demonstrate LRP1 as crucial host entry factor for PMT intoxication by acting as its primary cell surface receptor.

The role of the 5' sensing function of ribonuclease E in cyanobacteria.

RNA degradation is critical for synchronising gene expression with changing conditions in prokaryotic and eukaryotic organisms. In bacteria, the preference of the central ribonucleases RNase E, RNase J and RNase Y for 5'-monophosphorylated RNAs is considered important for RNA degradation. For RNase E, the underlying mechanism is termed 5' sensing, contrasting to the alternative 'direct entry' mode, which is independent of monophosphorylated 5' ends. Cyanobacteria, such as Synechocystis sp. PCC 6803 (Synechocystis), encode RNase E and RNase J homologues. Here, we constructed a Synechocystis strain lacking the 5' sensing function of RNase E and mapped on a transcriptome-wide level 283 5'-sensing-dependent cleavage sites. These included so far unknown targets such as mRNAs encoding proteins related to energy metabolism and carbon fixation. The 5' sensing function of cyanobacterial RNase E is important for the maturation of rRNA and several tRNAs, including tRNAGluUUC. This tRNA activates glutamate for tetrapyrrole biosynthesis in plant chloroplasts and in most prokaryotes. Furthermore, we found that increased RNase activities lead to a higher copy number of the major Synechocystis plasmids pSYSA and pSYSM. These results provide a first step towards understanding the importance of the different target mechanisms of RNase E outside Escherichia coli.

Transformer-based tool recommendation system in Galaxy.

BACKGROUND: Galaxy is a web-based open-source platform for scientific analyses. Researchers use thousands of high-quality tools and workflows for their respective analyses in Galaxy. Tool recommender system predicts a collection of tools that can be used to extend an analysis. In this work, a tool recommender system is developed by training a transformer on workflows available on Galaxy Europe and its performance is compared to other neural networks such as recurrent, convolutional and dense neural networks. RESULTS: The transformer neural network achieves two times faster convergence, has significantly lower model usage (model reconstruction and prediction) time and shows a better generalisation that goes beyond training workflows than the older tool recommender system created using RNN in Galaxy. In addition, the transformer also outperforms CNN and DNN on several key indicators. It achieves a faster convergence time, lower model usage time, and higher quality tool recommendations than CNN. Compared to DNN, it converges faster to a higher precision@k metric (approximately 0.98 by transformer compared to approximately 0.9 by DNN) and shows higher quality tool recommendations. CONCLUSION: Our work shows a novel usage of transformers to recommend tools for extending scientific workflows. A more robust tool recommendation model, created using a transformer, having significantly lower usage time than RNN and CNN, higher precision@k than DNN, and higher quality tool recommendations than all three neural networks, will benefit researchers in creating scientifically significant workflows and exploratory data analysis in Galaxy. Additionally, the ability to train faster than all three neural networks imparts more scalability for training on larger datasets consisting of millions of tool sequences. Open-source scripts to create the recommendation model are available under MIT licence at https://github.com/anuprulez/galaxy_tool_recommendation_transformers.

Peakhood: individual site context extraction for CLIP-seq peak regions.

MOTIVATION: CLIP-seq is by far the most widely used method to determine transcriptome-wide binding sites of RNA-binding proteins (RBPs). The binding site locations are identified from CLIP-seq read data by tools termed peak callers. Many RBPs bind to a spliced RNA (i.e. transcript) context, but all currently available peak callers only consider and report the genomic context. To accurately model protein binding behavior, a tool is needed for the individual context assignment to CLIP-seq peak regions. RESULTS: Here we present Peakhood, the first tool that utilizes CLIP-seq peak regions identified by peak callers, in tandem with CLIP-seq read information and genomic annotations, to determine which context applies, individually for each peak region. For sites assigned to transcript context, it further determines the most likely splice variant, and merges results for any number of datasets to obtain a comprehensive collection of transcript context binding sites. AVAILABILITY AND IMPLEMENTATION: Peakhood is freely available under MIT license at: https://github.com/BackofenLab/Peakhood. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

CRISPRtracrRNA: robust approach for CRISPR tracrRNA detection.

MOTIVATION: The CRISPR-Cas9 system is a Type II CRISPR system that has rapidly become the most versatile and widespread tool for genome engineering. It consists of two components, the Cas9 effector protein, and a single guide RNA that combines the spacer (for identifying the target) with the tracrRNA, a trans-activating small RNA required for both crRNA maturation and interference. While there are well-established methods for screening Cas effector proteins and CRISPR arrays, the detection of tracrRNA remains the bottleneck in detecting Class 2 CRISPR systems. RESULTS: We introduce a new pipeline CRISPRtracrRNA for screening and evaluation of tracrRNA candidates in genomes. This pipeline combines evidence from different components of the Cas9-sgRNA complex. The core is a newly developed structural model via covariance models from a sequence-structure alignment of experimentally validated tracrRNAs. As additional evidence, we determine the terminator signal (required for the tracrRNA transcription) and the RNA-RNA interaction between the CRISPR array repeat and the 5'-part of the tracrRNA. Repeats are detected via an ML-based approach (CRISPRidenify). Providing further evidence, we detect the cassette containing the Cas9 (Type II CRISPR systems) and Cas12 (Type V CRISPR systems) effector protein. Our tool is the first for detecting tracrRNA for Type V systems. AVAILABILITY AND IMPLEMENTATION: The implementation of the CRISPRtracrRNA is available on GitHub upon requesting the access permission, (https://github.com/BackofenLab/CRISPRtracrRNA). Data generated in this study can be obtained upon request to the corresponding person: Rolf Backofen (backofen@informatik.uni-freiburg.de). SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

DHX9 suppresses RNA processing defects originating from the Alu invasion of the human genome.

Transposable elements are viewed as 'selfish genetic elements', yet they contribute to gene regulation and genome evolution in diverse ways. More than half of the human genome consists of transposable elements. Alu elements belong to the short interspersed nuclear element (SINE) family of repetitive elements, and with over 1 million insertions they make up more than 10% of the human genome. Despite their abundance and the potential evolutionary advantages they confer, Alu elements can be mutagenic to the host as they can act as splice acceptors, inhibit translation of mRNAs and cause genomic instability. Alu elements are the main targets of the RNA-editing enzyme ADAR and the formation of Alu exons is suppressed by the nuclear ribonucleoprotein HNRNPC, but the broad effect of massive secondary structures formed by inverted-repeat Alu elements on RNA processing in the nucleus remains unknown. Here we show that DHX9, an abundant nuclear RNA helicase, binds specifically to inverted-repeat Alu elements that are transcribed as parts of genes. Loss of DHX9 leads to an increase in the number of circular-RNA-producing genes and amount of circular RNAs, translational repression of reporters containing inverted-repeat Alu elements, and transcriptional rewiring (the creation of mostly nonsensical novel connections between exons) of susceptible loci. Biochemical purifications of DHX9 identify the interferon-inducible isoform of ADAR (p150), but not the constitutively expressed ADAR isoform (p110), as an RNA-independent interaction partner. Co-depletion of ADAR and DHX9 augments the double-stranded RNA accumulation defects, leading to increased circular RNA production, revealing a functional link between these two enzymes. Our work uncovers an evolutionarily conserved function of DHX9. We propose that it acts as a nuclear RNA resolvase that neutralizes the immediate threat posed by transposon insertions and allows these elements to evolve as tools for the post-transcriptional regulation of gene expression.

CRISPRidentify: identification of CRISPR arrays using machine learning approach.

CRISPR-Cas are adaptive immune systems that degrade foreign genetic elements in archaea and bacteria. In carrying out their immune functions, CRISPR-Cas systems heavily rely on RNA components. These CRISPR (cr) RNAs are repeat-spacer units that are produced by processing of pre-crRNA, the transcript of CRISPR arrays, and guide Cas protein(s) to the cognate invading nucleic acids, enabling their destruction. Several bioinformatics tools have been developed to detect CRISPR arrays based solely on DNA sequences, but all these tools employ the same strategy of looking for repetitive patterns, which might correspond to CRISPR array repeats. The identified patterns are evaluated using a fixed, built-in scoring function, and arrays exceeding a cut-off value are reported. Here, we instead introduce a data-driven approach that uses machine learning to detect and differentiate true CRISPR arrays from false ones based on several features. Our CRISPR detection tool, CRISPRidentify, performs three steps: detection, feature extraction and classification based on manually curated sets of positive and negative examples of CRISPR arrays. The identified CRISPR arrays are then reported to the user accompanied by detailed annotation. We demonstrate that our approach identifies not only previously detected CRISPR arrays, but also CRISPR array candidates not detected by other tools. Compared to other methods, our tool has a drastically reduced false positive rate. In contrast to the existing tools, our approach not only provides the user with the basic statistics on the identified CRISPR arrays but also produces a certainty score as a practical measure of the likelihood that a given genomic region is a CRISPR array.

Transcriptome-wide in vivo mapping of cleavage sites for the compact cyanobacterial ribonuclease E reveals insights into its function and substrate recognition.

Ribonucleases are crucial enzymes in RNA metabolism and post-transcriptional regulatory processes in bacteria. Cyanobacteria encode the two essential ribonucleases RNase E and RNase J. Cyanobacterial RNase E is shorter than homologues in other groups of bacteria and lacks both the chloroplast-specific N-terminal extension as well as the C-terminal domain typical for RNase E of enterobacteria. In order to investigate the function of RNase E in the model cyanobacterium Synechocystis sp. PCC 6803, we engineered a temperature-sensitive RNase E mutant by introducing two site-specific mutations, I65F and the spontaneously occurred V94A. This enabled us to perform RNA-seq after the transient inactivation of RNase E by a temperature shift (TIER-seq) and to map 1472 RNase-E-dependent cleavage sites. We inferred a dominating cleavage signature consisting of an adenine at the -3 and a uridine at the +2 position within a single-stranded segment of the RNA. The data identified mRNAs likely regulated jointly by RNase E and an sRNA and potential 3' end-derived sRNAs. Our findings substantiate the pivotal role of RNase E in post-transcriptional regulation and suggest the redundant or concerted action of RNase E and RNase J in cyanobacteria.

CRISPRloci: comprehensive and accurate annotation of CRISPR-Cas systems.

CRISPR-Cas systems are adaptive immune systems in prokaryotes, providing resistance against invading viruses and plasmids. The identification of CRISPR loci is currently a non-standardized, ambiguous process, requiring the manual combination of multiple tools, where existing tools detect only parts of the CRISPR-systems, and lack quality control, annotation and assessment capabilities of the detected CRISPR loci. Our CRISPRloci server provides the first resource for the prediction and assessment of all possible CRISPR loci. The server integrates a series of advanced Machine Learning tools within a seamless web interface featuring: (i) prediction of all CRISPR arrays in the correct orientation; (ii) definition of CRISPR leaders for each locus; and (iii) annotation of cas genes and their unambiguous classification. As a result, CRISPRloci is able to accurately determine the CRISPR array and associated information, such as: the Cas subtypes; cassette boundaries; accuracy of the repeat structure, orientation and leader sequence; virus-host interactions; self-targeting; as well as the annotation of cas genes, all of which have been missing from existing tools. This annotation is presented in an interactive interface, making it easy for scientists to gain an overview of the CRISPR system in their organism of interest. Predictions are also rendered in GFF format, enabling in-depth genome browser inspection. In summary, CRISPRloci constitutes a full suite for CRISPR-Cas system characterization that offers annotation quality previously available only after manual inspection.

Meta-Analysis of Mechano-Sensitive Ion Channels in Human Hearts: Chamber- and Disease-Preferential mRNA Expression.

The cardiac cell mechanical environment changes on a beat-by-beat basis as well as in the course of various cardiac diseases. Cells sense and respond to mechanical cues via specialized mechano-sensors initiating adaptive signaling cascades. With the aim of revealing new candidates underlying mechano-transduction relevant to cardiac diseases, we investigated mechano-sensitive ion channels (MSC) in human hearts for their chamber- and disease-preferential mRNA expression. Based on a meta-analysis of RNA sequencing studies, we compared the mRNA expression levels of MSC in human atrial and ventricular tissue samples from transplant donor hearts (no cardiac disease), and from patients in sinus rhythm (underlying diseases: heart failure, coronary artery disease, heart valve disease) or with atrial fibrillation. Our results suggest that a number of MSC genes are expressed chamber preferentially, e.g., CHRNE in the atria (compared to the ventricles), TRPV4 in the right atrium (compared to the left atrium), CACNA1B and KCNMB1 in the left atrium (compared to the right atrium), as well as KCNK2 and KCNJ2 in ventricles (compared to the atria). Furthermore, 15 MSC genes are differentially expressed in cardiac disease, out of which SCN9A (lower expressed in heart failure compared to donor tissue) and KCNQ5 (lower expressed in atrial fibrillation compared to sinus rhythm) show a more than twofold difference, indicative of possible functional relevance. Thus, we provide an overview of cardiac MSC mRNA expression in the four cardiac chambers from patients with different cardiac diseases. We suggest that the observed differences in MSC mRNA expression may identify candidates involved in altered mechano-transduction in the respective diseases.

Scool: a new data storage format for single-cell Hi-C data.

MOTIVATION: Single-cell Hi-C research currently lacks an efficient, easy to use and shareable data storage format. Recent studies have used a variety of sub-optimal solutions: publishing raw data only, text-based interaction matrices, or reusing established Hi-C storage formats for single interaction matrices. These approaches are storage and pre-processing intensive, require long labour time and are often error-prone. RESULTS: The single-cell cooler file format (scool) provides an efficient, user-friendly and storage-saving approach for single-cell Hi-C data. It is a flavour of the established cooler format and guarantees stable API support. AVAILABILITY AND IMPLEMENTATION: The single-cell cooler format is part of the cooler file format as of API version 0.8.9. It is available via pip, conda and github: https://github.com/mirnylab/cooler. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.