Multimodal integration combines information from different sources or modalities to gain a more comprehensive understanding of a phenomenon. The challenges in multi-omics data analysis lie in the complexity, high dimensionality, and heterogeneity of the data, which demands sophisticated computational tools and visualization methods for proper interpretation and visualization of multi-omics data. In this paper, we propose a novel method, termed Orthogonal Multimodality Integration and Clustering (OMIC), for analyzing CITE-seq. Our approach enables researchers to integrate multiple sources of information while accounting for the dependence among them. We demonstrate the effectiveness of our approach using CITE-seq data sets for cell clustering. Our results show that our approach outperforms existing methods in terms of accuracy, computational efficiency, and interpretability. We conclude that our proposed OMIC method provides a powerful tool for multimodal data analysis that greatly improves the feasibility and reliability of integrated data.

MOTIVATION: Genome-wide profiling of transcription factor binding and chromatin states is a widely-used approach for mechanistic understanding of gene regulation. Recent technology development has enabled such profiling at single-cell resolution. However, an end-to-end computational pipeline for analyzing such data is still lacking. RESULTS: Here, we have developed a flexible pipeline for analysis and visualization of single-cell CUT&Tag and CUT&RUN data, which provides functions for sequence alignment, quality control, dimensionality reduction, cell clustering, data aggregation and visualization. Furthermore, it is also seamlessly integrated with the functions in original CUT&RUNTools for population-level analyses. As such, this provides a valuable toolbox for the community. AVAILABILITY AND IMPLEMENTATION: https://github.com/fl-yu/CUT-RUNTools-2.0. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

BACKGROUND: With the rapid development of single-cell RNA sequencing technology, it is possible to dissect cell-type composition at high resolution. A number of methods have been developed with the purpose to identify rare cell types. However, existing methods are still not scalable to large datasets, limiting their utility. To overcome this limitation, we present a new software package, called GiniClust3, which is an extension of GiniClust2 and significantly faster and memory-efficient than previous versions. RESULTS: Using GiniClust3, it only takes about 7 h to identify both common and rare cell clusters from a dataset that contains more than one million cells. Cell type mapping and perturbation analyses show that GiniClust3 could robustly identify cell clusters. CONCLUSIONS: Taken together, these results suggest that GiniClust3 is a powerful tool to identify both common and rare cell population and can handle large dataset. GiniCluster3 is implemented in the open-source python package and available at https://github.com/rdong08/GiniClust3.

BACKGROUND: Single-cell RNA-sequencing technologies provide a powerful tool for systematic dissection of cellular heterogeneity. However, the prevalence of dropout events imposes complications during data analysis and, despite numerous efforts from the community, this challenge has yet to be solved. RESULTS: Here we present a computational method, called RESCUE, to mitigate the dropout problem by imputing gene expression levels using information from other cells with similar patterns. Unlike existing methods, we use an ensemble-based approach to minimize the feature selection bias on imputation. By comparative analysis of simulated and real single-cell RNA-seq datasets, we show that RESCUE outperforms existing methods in terms of imputation accuracy which leads to more precise cell-type identification. CONCLUSIONS: Taken together, these results suggest that RESCUE is a useful tool for mitigating dropouts in single-cell RNA-seq data. RESCUE is implemented in R and available at https://github.com/seasamgo/rescue .

Motivation: With the increasing amount of genomic and epigenomic data in the public domain, a pressing challenge is to integrate these data to investigate the role of epigenetic mechanisms in regulating gene expression and maintenance of cell-identity. To this end, we have implemented a computational pipeline to systematically study epigenetic variability and uncover regulatory DNA sequences. Results: Haystack is a bioinformatics pipeline to identify hotspots of epigenetic variability across different cell-types, cell-type specific cis-regulatory elements, and associated transcription factors. Haystack is generally applicable to any epigenetic mark and provides an important tool to investigate the mechanisms underlying epigenetic switches during development. This software is accompanied by a set of precomputed tracks, which may be used as a valuable resource for functional annotation of the human genome. Availability and implementation: The Haystack pipeline is implemented as an open-source, multiplatform, Python package called haystack_bio freely available at https://github.com/pinellolab/haystack_bio. Contact: [email protected] or [email protected]. Supplementary information: Supplementary data are available at Bioinformatics online.

BACKGROUND: We recently identified two robust ovarian cancer subtypes, defined by the expression of genes involved in angiogenesis, with significant differences in clinical outcome. To identify potential regulatory mechanisms that distinguish the subtypes we applied PANDA, a method that uses an integrative approach to model information flow in gene regulatory networks. RESULTS: We find distinct differences between networks that are active in the angiogenic and non-angiogenic subtypes, largely defined by a set of key transcription factors that, although previously reported to play a role in angiogenesis, are not strongly differentially-expressed between the subtypes. Our network analysis indicates that these factors are involved in the activation (or repression) of different genes in the two subtypes, resulting in differential expression of their network targets. Mechanisms mediating differences between subtypes include a previously unrecognized pro-angiogenic role for increased genome-wide DNA methylation and complex patterns of combinatorial regulation. CONCLUSIONS: The models we develop require a shift in our interpretation of the driving factors in biological networks away from the genes themselves and toward their interactions. The observed regulatory changes between subtypes suggest therapeutic interventions that may help in the treatment of ovarian cancer.

Epigenetic mechanisms play an important role in the regulation of cell type-specific gene activities, yet how epigenetic patterns are established and maintained remains poorly understood. Recent studies have supported a role of DNA sequences in recruitment of epigenetic regulators. Alignment-free methods have been applied to identify distinct sequence features that are associated with epigenetic patterns and to predict epigenomic profiles. Here, we review recent advances in such applications, including the methods to map DNA sequence to feature space, sequence comparison and prediction models. Computational studies using these methods have provided important insights into the epigenetic regulatory mechanisms.

BACKGROUND: Genome-wide mapping of protein-DNA interactions has been widely used to investigate biological functions of the genome. An important question is to what extent such interactions are regulated at the DNA sequence level. However, current investigation is hampered by the lack of computational methods for systematic evaluating sequence specificity. RESULTS: We present a simple, unbiased quantitative measure for DNA sequence specificity called the Motif Independent Measure (MIM). By analyzing both simulated and real experimental data, we found that the MIM measure can be used to detect sequence specificity independent of presence of transcription factor (TF) binding motifs. We also found that the level of specificity associated with H3K4me1 target sequences is highly cell-type specific and highest in embryonic stem (ES) cells. We predicted H3K4me1 target sequences by using the N- score model and found that the prediction accuracy is indeed high in ES cells.The software to compute the MIM is freely available at: https://github.com/lucapinello/mim. CONCLUSIONS: Our method provides a unified framework for quantifying DNA sequence specificity and serves as a guide for development of sequence-based prediction models.

BACKGROUND: Epigenetics is an important layer of transcriptional control necessary for cell-type specific gene regulation. Recent studies have shown significant epigenetic patterns associated with developmental stages and diseases. However, previous studies have been mostly limited to focal epigenetic patterns, whereas methods for analyzing large-scale organizations are still lacking. RESULTS: We developed a hidden Markov model (HMM) approach for detecting the types and locations of epigenetic domains from multiple histone modifications. We used this method to analyze a published ChIP-seq dataset of five histone modification marks (H3K4me2, H3K4me3, H3K27me3, H3K9me3, and H3K36me3) in mouse embryonic stem (ES) cells. We identified three types of domains, corresponding to active, non-active, and null states. In total, our three-state HMM identified 258 domains in the mouse genome containing 9.6 genes on average. These domains were validated by a number of criteria. The largest domains correspond to olfactory receptor (OR) gene clusters. Each Hox gene cluster also forms a separate epigenetic domain. We found that each type of domain is associated with distinct biological functions and structural changes during early cell differentiation. CONCLUSIONS: The HMM approach successfully detects domains of consistent epigenetic patterns from ChIP-seq data, providing new insights into the role of epigenetics in long-range gene regulation.

We developed a computational approach to detect multi- gene domains with coherent epigenetic patterns. We applied this method to analyse a published ChIP-seq dataset for five different histone modification marks (H3K4me2, H3K4me3, H3K27me3, H3K9me3, H3K36me3) in mouse embryonic stem cells. We first obtained a 5-dimenisinal score for all known genes based on average modification activity in select regions. Then, with hidden Markov models and corresponding algorithms, we were able to determine the most probable domain status of each gene. We find that a three-state hidden Markov model can best describe the data, where the states correspond to active, inactive, and null domains. This model predicts 339 significantly large multi-gene domains, including known domains such as the olfactory receptor (OR) gene clusters, but also previously uncharacterized domains (Figure 1 ). We also noted less histone modification variability within each of our domains when compared to randomly selected boundaries. We further validated our predictions against gene expression and Gene Ontology data and found our domains were functionally relevant. Heatmaps of known gene clusters . The 35 gene region on Ch6 from Npy to 2410066E13Rik as depictured as a heatmap of histone modification and gene expression for (a) the ES cell line and (c) ES and NP cell lines. The 250 gene region on Ch7 from rt2a to Insc as depictured as a heatmap of histone modification and gene expression for (b) the ES cell line and (d) ES and NP cell lines. For all figures on the left (a and c), cluster state assignment is given in the first and second tracks and HMM state assignments are in the third and fourth tracks (red for active state, blue for non-active state, yellow for null state). For figures on the right (b and d), NP HMM state assignment is given in the first and second tracks and ES HMM state assignments are in the third and fourth tracks. Whether (black) or not (white) a gene is a respective gene cluster in shown in the bottom track in all figures. Our method provides a novel approach to analyse large-scale epigenetic patterns. As we continue to apply our method to other cell lines, we will provide important insight into the general structure, organization, and regulation of the mammalian genome.

The regulation of DNA accessibility through nucleosome positioning is important for transcription control. Computational models have been developed to predict genome-wide nucleosome positions from DNA sequences, but these models consider only nucleosome sequences, which may have limited their power. We developed a statistical multi-resolution approach to identify a sequence signature, called the N-score, that distinguishes nucleosome binding DNA from non-nucleosome DNA. This new approach has significantly improved the prediction accuracy. The sequence information is highly predictive for local nucleosome enrichment or depletion, whereas predictions of the exact positions are only modestly more accurate than a null model, suggesting the importance of other regulatory factors in fine-tuning the nucleosome positions. The N-score in promoter regions is negatively correlated with gene expression levels. Regulatory elements are enriched in low N-score regions. While our model is derived from yeast data, the N-score pattern computed from this model agrees well with recent high-resolution protein-binding data in human.

In eukaryotic cells, genomic DNAs wrap around beadlike molecules, called nucleosomes, so as to pack more compactly in the nucleus of the cell. The nucleosome is made up of four pairs of histone proteins (H2A, H2B, H3, and H4) who share a very similar structural motif. The positioning of nucleosomes as well as the modifications of various sites of histone proteins (such as acetylation) plays important but incompletely understood roles in gene regulation. We propose statistical models for predicting nucleosome positioning and histone modification patterns using only genomic sequence information. Computation models have been developed to predict genome-wide nucleosome positions from DNA sequences, but these models consider only nucleosome sequences, which may have limited their power. We developed a statistical multi-resolution approach to identify a sequence signature, called the N-score, that distinguishes nucleosome binding DNA from non-nucleosome DNA. The N-score is not sensitive to deletion of short DNA elements and can also be estimated reasonably accurately from coarse nucleosome positioning data. We found that the sequence information is highly predictive for local nucleosome enrichment or depletion, whereas the exact positions may be further fine-tuned by other regulatory factors. We observed that many characteristics of nucleosome positioning, such as the nucleosome depletion in the promoter regions, can be predicted accurately by the sequence information through N-scores. In addition to nucleosome positioning, histone acetylations are also important in directing gene regulation. A comprehensive understanding of the regulatory role of histone acetylation is difficult because many different histone acetylation patterns exist and their effects are confounded by other factors, such as the transcription factor binding sequence motif information and nucleosome occupancy. We analyzed recent genomewide histone acetylation data using a few complementary statistical models and tested the validity of a cumulative model in approximating the global regulatory effect of histone acetylation. Confounding effects due to transcription factor binding sequence information were estimated by using two independent motif-based algorithms followed by a variable selection method. Our analysis confirms that histone acetylation has a significant effect on transcription rates in addition to that attributable to upstream sequence motifs. Our model fits well with observed genome-wide data. Strikingly, including more complicated combinatorial effects does not improve the model's performance. Through an analysis of conditional independence, we found that H4 acetylation may not have significant direct impact on global gene expression.