Detection of DNA base modifications by deep recurrent neural network on Oxford Nanopore sequencing data

Methods in DNA methylation array dataset analysis: A review

Understanding the intricate relationships between gene expression levels and epigenetic modifications in a genome is crucial to comprehending the pathogenic mechanisms of many diseases. With the advancement of DNA Methylome Profiling techniques, the emphasis on identifying Differentially Methylated Regions (DMRs/DMGs) has become crucial for biomarker discovery, offering new insights into the etiology of illnesses. This review surveys the current state of computational tools/algorithms for the analysis of microarray-based DNA methylation profiling datasets, focusing on key concepts underlying the diagnostic/prognostic CpG site extraction. It addresses methodological frameworks, algorithms, and pipelines employed by various authors, serving as a roadmap to address challenges and understand changing trends in the methodologies for analyzing array-based DNA methylation profiling datasets derived from diseased genomes. Additionally, it highlights the importance of integrating gene expression and methylation datasets for accurate biomarker identification, explores prognostic prediction models, and discusses molecular subtyping for disease classification. The review also emphasizes the contributions of machine learning, neural networks, and data mining to enhance diagnostic workflow development, thereby improving accuracy, precision, and robustness.

Rockfish: A transformer-based model for accurate 5-methylcytosine prediction from nanopore sequencing

DNA methylation plays an important role in various biological processes, including cell differentiation, ageing, and cancer development. The most important methylation in mammals is 5-methylcytosine mostly occurring in the context of CpG dinucleotides. Sequencing methods such as whole-genome bisulfite sequencing successfully detect 5-methylcytosine DNA modifications. However, they suffer from the serious drawbacks of short read lengths and might introduce an amplification bias. Here we present Rockfish, a deep learning algorithm that significantly improves read-level 5-methylcytosine detection by using Nanopore sequencing. Rockfish is compared with other methods based on Nanopore sequencing on R9.4.1 and R10.4.1 datasets. There is an increase in the single-base accuracy and the F1 measure of up to 5 percentage points on R.9.4.1 datasets, and up to 0.82 percentage points on R10.4.1 datasets. Moreover, Rockfish shows a high correlation with whole-genome bisulfite sequencing, requires lower read depth, and achieves higher confidence in biologically important regions such as CpG-rich promoters while being computationally efficient. Its superior performance in human and mouse samples highlights its versatility for studying 5-methylcytosine methylation across varied organisms and diseases. Finally, its adaptable architecture ensures compatibility with new versions of pores and chemistry as well as modification types.

Targeting transposable elements in cancer: developments and opportunities

Transposable elements (TEs), comprising nearly 50% of the human genome, have transitioned from being perceived as "genomic junk" to key players in cancer progression. Contemporary research links TE regulatory disruptions with cancer development, underscoring their therapeutic potential. Advances in long-read sequencing, computational analytics, single-cell sequencing, proteomics, and CRISPR-Cas9 technologies have enriched our understanding of TEs' clinical implications, notably their impact on genome architecture, gene regulation, and evolutionary processes. In cancer, TEs, including long interspersed element-1 (LINE-1), Alus, and long terminal repeat (LTR) elements, demonstrate altered patterns, influencing both tumorigenic and tumor-suppressive mechanisms. TE-derived nucleic acids and tumor antigens play critical roles in tumor immunity, bridging innate and adaptive responses. Given their central role in oncology, TE-targeted therapies, particularly through reverse transcriptase inhibitors and epigenetic modulators, represent a novel avenue in cancer treatment. Combining these TE-focused strategies with existing chemotherapy or immunotherapy regimens could enhance efficacy and offer a new dimension in cancer treatment. This review delves into recent TE detection advancements, explores their multifaceted roles in tumorigenesis and immune regulation, discusses emerging diagnostic and therapeutic approaches centered on TEs, and anticipates future directions in cancer research.

MethPhaser: methylation-based long-read haplotype phasing of human genomes

The assignment of variants across haplotypes, phasing, is crucial for predicting the consequences, interaction, and inheritance of mutations and is a key step in improving our understanding of phenotype and disease. However, phasing is limited by read length and stretches of homozygosity along the genome. To overcome this limitation, we designed MethPhaser, a method that utilizes methylation signals from Oxford Nanopore Technologies to extend Single Nucleotide Variation (SNV)-based phasing. We demonstrate that haplotype-specific methylations extensively exist in Human genomes and the advent of long-read technologies enabled direct report of methylation signals. For ONT R9 and R10 cell line data, we increase the phase length N50 by 78%-151% at a phasing accuracy of 83.4-98.7% To assess the impact of tissue purity and random methylation signals due to inactivation, we also applied MethPhaser on blood samples from 4 patients, still showing improvements over SNV-only phasing. MethPhaser further improves phasing across HLA and multiple other medically relevant genes, improving our understanding of how mutations interact across multiple phenotypes. The concept of MethPhaser can also be extended to non-human diploid genomes. MethPhaser is available at https://github.com/treangenlab/methphaser .

Precision Basecalling of Single DNA Nucleotide from Overlapped Transmission Readouts with Machine Learning Aided Solid-State Nanogap

DNA sequencing with the quantum tunneling technique heralds a paradigm shift in genetic analysis, promising rapid and accurate identification for diverging applications ranging from personalized medicine to security issues. However, the widespread distribution of molecular conductance, conduction orbital alignment for resonant transport, and decoding crisscrossing conductance signals of isomorphic nucleotides have been persistent experimental hurdles for swift and precise identification. Herein, we have reported a machine learning (ML)-driven quantum tunneling study with solid-state model nanogap to determine nucleotides at single-base resolution. The optimized ML basecaller has demonstrated a high predictive basecalling accuracy of all four nucleotides from seven distinct data pools, each containing complex transmission readouts of their different dynamic conformations. ML classification of quaternary, ternary, and binary nucleotide combinations is also performed with high precision, sensitivity, and F1 score. ML explainability unravels the evidence of how extracted normalized features within overlapped nucleotide signals contribute to classification improvement. Moreover, electronic fingerprints, conductance sensitivity, and current readout analysis of nucleotides have promised practical applicability with significant sensitivity and distinguishability. Through this ML approach, our study pushes the boundaries of quantum sequencing by highlighting the effectiveness of single nucleotide basecalling with promising implications for advancing genomics and molecular diagnostics.

Effective training of nanopore callers for epigenetic marks with limited labelled data

Nanopore sequencing platforms combined with supervised machine learning (ML) have been effective at detecting base modifications in DNA such as 5-methylcytosine (5mC) and N6-methyladenine (6mA). These ML-based nanopore callers have typically been trained on data that span all modifications on all possible DNA [Formula: see text]-mer backgrounds-a complete training dataset. However, as nanopore technology is pushed to more and more epigenetic modifications, such complete training data will not be feasible to obtain. Nanopore calling has historically been performed with hidden Markov models (HMMs) that cannot make successful calls for [Formula: see text]-mer contexts not seen during training because of their independent emission distributions. However, deep neural networks (DNNs), which share parameters across contexts, are increasingly being used as callers, often outperforming their HMM cousins. It stands to reason that a DNN approach should be able to better generalize to unseen [Formula: see text]-mer contexts. Indeed, herein we demonstrate that a common DNN approach (DeepSignal) outperforms a common HMM approach (Nanopolish) in the incomplete data setting. Furthermore, we propose a novel hybrid HMM-DNN approach, amortized-HMM, that outperforms both the pure HMM and DNN approaches on 5mC calling when the training data are incomplete. This type of approach is expected to be useful for calling other base modifications such as 5-hydroxymethylcytosine and for the simultaneous calling of different modifications, settings in which complete training data are not likely to be available.

DNA modifications: Biomarkers for the exposome?

The concept of the exposome is the encompassing of all the environmental exposures, both exogenous and endogenous, across the life course. Many, if not all, of these exposures can result in the generation of reactive species, and/or the modulation of cellular processes, that can lead to a breadth of modifications of DNA, the nature of which may be used to infer their origin. Because of their role in cell function, such modifications have been associated with various major human diseases, including cancer, and so their assessment is crucial. Historically, most methods have been able to only measure one or a few DNA modifications at a time, limiting the information available. With the development of DNA adductomics, which aims to determine the totality of DNA modifications, a far more comprehensive picture of the DNA adduct burden can be gained. Importantly, DNA adductomics can facilitate a "top-down" investigative approach whereby patterns of adducts may be used to trace and identify the originating exposure source. This, together with other 'omic approaches, represents a major tool for unraveling the complexities of the exposome and hence allow a better a understanding of the environmental origins of disease.

Can long-read sequencing tackle the barriers, which the next-generation could not? A review

The large-scale heterogeneity of genetic diseases necessitated the deeper examination of nucleotide sequence alterations enhancing the discovery of new targeted drug attack points. The appearance of new sequencing techniques was essential to get more interpretable genomic data. In contrast to the previous short-reads, longer lengths can provide a better insight into the potential health threatening genetic abnormalities. Long-reads offer more accurate variant identification and genome assembly methods, indicating advances in nucleotide deflect-related studies. In this review, we introduce the historical background of sequencing technologies and show their benefits and limits, as well. Furthermore, we highlight the differences between short- and long-read approaches, including their unique advances and difficulties in methodologies and evaluation. Additionally, we provide a detailed description of the corresponding bioinformatics and the current applications.

Population epigenetics: DNA methylation in the plant omics era

DNA methylation plays an important role in many biological processes. The mechanisms underlying the establishment and maintenance of DNA methylation are well understood thanks to decades of research using DNA methylation mutants, primarily in Arabidopsis (Arabidopsis thaliana) accession Col-0. Recent genome-wide association studies (GWASs) using the methylomes of natural accessions have uncovered a complex and distinct genetic basis of variation in DNA methylation at the population level. Sequencing following bisulfite treatment has served as an excellent method for quantifying DNA methylation. Unlike studies focusing on specific accessions with reference genomes, population-scale methylome research often requires an additional round of sequencing beyond obtaining genome assemblies or genetic variations from whole-genome sequencing data, which can be cost prohibitive. Here, we provide an overview of recently developed bisulfite-free methods for quantifying methylation and cost-effective approaches for the simultaneous detection of genetic and epigenetic information. We also discuss the plasticity of DNA methylation in a specific Arabidopsis accession, the contribution of DNA methylation to plant adaptation, and the genetic determinants of variation in DNA methylation in natural populations. The recently developed technology and knowledge will greatly benefit future studies in population epigenomes.

Protocol to isolate nuclei from Chlamydomonas reinhardtii for ATAC sequencing

The isolation of sufficient amounts of intact nuclei is essential to obtain high-resolution maps of chromatin accessibility via assay for transposase-accessible chromatin using sequencing (ATAC-seq). Here, we present a protocol for tag-free isolation of nuclei from both cell walled and cell wall-deficient strains of the green model alga Chlamydomonas reinhardtii at a suitable quality for ATAC-seq. We describe steps for nuclei isolation, quantification, and downstream ATAC-seq. This protocol is optimized to shorten the time of isolation and quantification of nuclei.

A comparison of methods for detecting DNA methylation from long-read sequencing of human genomes

BACKGROUND

Long-read sequencing can enable the detection of base modifications, such as CpG methylation, in single molecules of DNA. The most commonly used methods for long-read sequencing are nanopore developed by Oxford Nanopore Technologies (ONT) and single molecule real-time (SMRT) sequencing developed by Pacific Bioscience (PacBio). In this study, we systematically compare the performance of CpG methylation detection from long-read sequencing.

RESULTS

We demonstrate that CpG methylation detection from 7179 nanopore-sequenced DNA samples is highly accurate and consistent with 132 oxidative bisulfite-sequenced (oxBS) samples, isolated from the same blood draws. We introduce quality filters for CpGs that further enhance the accuracy of CpG methylation detection from nanopore-sequenced DNA, while removing at most 30% of CpGs. We evaluate the per-site performance of CpG methylation detection across different genomic features and CpG methylation rates and demonstrate how the latest R10.4 flowcell chemistry and base-calling algorithms improve methylation detection from nanopore sequencing. Additionally, we show how the methylation detection of 50 SMRT-sequenced genomes compares to nanopore sequencing and oxBS.

CONCLUSIONS

This study provides the first systematic comparison of CpG methylation detection tools for long-read sequencing methods. We compare two commonly used computational methods for the detection of CpG methylation in a large number of nanopore genomes, including samples sequenced using the latest R10.4 nanopore flowcell chemistry and 50 SMRT sequenced samples. We provide insights into the strengths and limitations of each sequencing method as well as recommendations for standardization and evaluation of tools designed for genome-scale modified base detection using long-read sequencing.

Detecting DNA hydroxymethylation: exploring its role in genome regulation

DNA methylation is one of the most extensively studied epigenetic regulatory mechanisms, known to play crucial roles in various organisms. It has been implicated in the regulation of gene expression and chromatin changes, ranging from global alterations during cell state transitions to locus-specific modifications. 5-hydroxymethylcytosine (5hmC) is produced by a major oxidation, from 5-methylcytosine (5mC), catalyzed by the ten-eleven translocation (TET) enzymes, and is gradually being recognized for its significant role in genome regulation. With the development of state-of-the-art experimental techniques, it has become possible to detect and distinguish 5mC and 5hmC at base resolution. Various techniques have evolved, encompassing chemical and enzymatic approaches, as well as thirdgeneration sequencing techniques. These advancements have paved the way for a thorough exploration of the role of 5hmC across a diverse array of cell types, from embryonic stem cells (ESCs) to various differentiated cells. This review aims to comprehensively report on recent techniques and discuss the emerging roles of 5hmC. [BMB Reports 2024; 57(3): 135-142].

A signal processing and deep learning framework for methylation detection using Oxford Nanopore sequencing

Oxford Nanopore sequencing can detect DNA methylations from ionic current signal of single molecules, offering a unique advantage over conventional methods. Additionally, adaptive sampling, a software-controlled enrichment method for targeted sequencing, allows reduced representation methylation sequencing that can be applied to CpG islands or imprinted regions. Here we present DeepMod2, a comprehensive deep-learning framework for methylation detection using ionic current signal from Nanopore sequencing. DeepMod2 implements both a bidirectional long short-term memory (BiLSTM) model and a Transformer model and can analyze POD5 and FAST5 signal files generated on R9 and R10 flowcells. Additionally, DeepMod2 can run efficiently on central processing unit (CPU) through model pruning and can infer epihaplotypes or haplotype-specific methylation calls from phased reads. We use multiple publicly available and newly generated datasets to evaluate the performance of DeepMod2 under varying scenarios. DeepMod2 has comparable performance to Guppy and Dorado, which are the current state-of-the-art methods from Oxford Nanopore Technologies that remain closed-source. Moreover, we show a high correlation (r = 0.96) between reduced representation and whole-genome Nanopore sequencing. In summary, DeepMod2 is an open-source tool that enables fast and accurate DNA methylation detection from whole-genome or adaptive sequencing data on a diverse range of flowcell types.

Extraction and selection of high-molecular-weight DNA for long-read sequencing from Chlamydomonas reinhardtii

Recent advances in long-read sequencing technologies have enabled the complete assembly of eukaryotic genomes from telomere to telomere by allowing repeated regions to be fully sequenced and assembled, thus filling the gaps left by previous short-read sequencing methods. Furthermore, long-read sequencing can also help characterizing structural variants, with applications in the fields of genome evolution or cancer genomics. For many organisms, the main bottleneck to sequence long reads remains the lack of robust methods to obtain high-molecular-weight (HMW) DNA. For this purpose, we developed an optimized protocol to extract DNA suitable for long-read sequencing from the unicellular green alga Chlamydomonas reinhardtii, based on CTAB/phenol extraction followed by a size selection step for long DNA molecules. We provide validation results for the extraction protocol, as well as statistics obtained with Oxford Nanopore Technologies sequencing.

T-S2Inet: Transformer-based sequence-to-image network for accurate nanopore sequence recognition

MOTIVATION

Nanopore sequencing is a new macromolecular recognition and perception technology that enables high-throughput sequencing of DNA, RNA, even protein molecules. The sequences generated by nanopore sequencing span a large time frame, and the labor and time costs incurred by traditional analysis methods are substantial. Recently, research on nanopore data analysis using machine learning algorithms has gained unceasing momentum, but there is often a significant gap between traditional and deep learning methods in terms of classification results. To analyze nanopore data using deep learning technologies, measures such as sequence completion and sequence transformation can be employed. However, these technologies do not preserve the local features of the sequences. To address this issue, we propose a sequence-to-image (S2I) module that transforms sequences of unequal length into images. Additionally, we propose the Transformer-based T-S2Inet model to capture the important information and improve the classification accuracy.

RESULTS

Quantitative and qualitative analysis shows that the experimental results have an improvement of around 2% in accuracy compared to previous methods. The proposed method is adaptable to other nanopore platforms, such as the Oxford nanopore. It is worth noting that the proposed method not only aims to achieve the most advanced performance, but also provides a general idea for the analysis of nanopore sequences of unequal length.

AVAILABILITY AND IMPLEMENTATION

The main program is available at https://github.com/guanxiaoyu11/S2Inet.

Chemical Methods to Identify Epigenetic Modifications in Cytosine Bases

Chemical modifications to Cytosine bases are among the most studied epigenetic markers and their detection in the human genome plays a crucial role in gaining more insights about gene regulation, prognosis of genetic disorders and unraveling genetic inheritance patterns. The Cytosine methylated at the 5th position and oxidized derivatives thereof generated in the demethylation pathways, perform separate and unique epigenetic functions in an organism. As the presence of various Cytosine modifications is associated with diverse diseases, including cancer, there has been a strong focus on developing methods, both chemical and alternative approaches, capable of detecting these modifications at a single-base resolution across the entire genome. In this comprehensive review, we aim to consolidate the various chemical methods and understanding their chemistry that have been established to date for the detection of various Cytosine modifications.

Toward Explainable Artificial Intelligence for Precision Pathology

The rapid development of precision medicine in recent years has started to challenge diagnostic pathology with respect to its ability to analyze histological images and increasingly large molecular profiling data in a quantitative, integrative, and standardized way. Artificial intelligence (AI) and, more precisely, deep learning technologies have recently demonstrated the potential to facilitate complex data analysis tasks, including clinical, histological, and molecular data for disease classification; tissue biomarker quantification; and clinical outcome prediction. This review provides a general introduction to AI and describes recent developments with a focus on applications in diagnostic pathology and beyond. We explain limitations including the black-box character of conventional AI and describe solutions to make machine learning decisions more transparent with so-called explainable AI. The purpose of the review is to foster a mutual understanding of both the biomedical and the AI side. To that end, in addition to providing an overview of the relevant foundations in pathology and machine learning, we present worked-through examples for a better practical understanding of what AI can achieve and how it should be done.

Epigenomic Characterization of Lymphoid Neoplasms

Lymphoid neoplasms represent a heterogeneous group of disease entities and subtypes with markedly different molecular and clinical features. Beyond genetic alterations, lymphoid tumors also show widespread epigenomic changes. These severely affect the levels and distribution of DNA methylation, histone modifications, chromatin accessibility, and three-dimensional genome interactions. DNA methylation stands out as a tracer of cell identity and memory, as B cell neoplasms show epigenetic imprints of their cellular origin and proliferative history, which can be quantified by an epigenetic mitotic clock. Chromatin-associated marks are informative to uncover altered regulatory regions and transcription factor networks contributing to the development of distinct lymphoid tumors. Tumor-intrinsic epigenetic and genetic aberrations cooperate and interact with microenvironmental cells to shape the transcriptome at different phases of lymphoma evolution, and intraclonal heterogeneity can now be characterized by single-cell profiling. Finally, epigenetics offers multiple clinical applications, including powerful diagnostic and prognostic biomarkers as well as therapeutic targets.

Nutrient inputs and social metabolic control of T cell fate

Cells in multicellular organisms experience diverse neighbors, signals, and evolving physical environments that drive functional and metabolic demands. To maintain proper development and homeostasis while avoiding inappropriate cell proliferation or death, individual cells interact with their neighbors via "social" cues to share and partition available nutrients. Metabolic signals also contribute to cell fate by providing biochemical links between cell-extrinsic signals and available resources. In addition to metabolic checkpoints that sense nutrients and directly supply molecular intermediates for biosynthetic pathways, many metabolites directly signal or provide the basis for post-translational modifications of target proteins and chromatin. In this review, we survey the landscape of T cell nutrient sensing and metabolic signaling that supports proper immunity while avoiding immunodeficiency or autoimmunity. The integration of cell-extrinsic microenvironmental cues with cell-intrinsic metabolic signaling provides a social metabolic control model to integrate cell signaling, metabolism, and fate.

An effective strategy for assembling the sex-limited chromosome

BACKGROUND

Most currently available reference genomes lack the sequence map of sex-limited (such as Y and W) chromosomes, which results in incomplete assemblies that hinder further research on sex chromosomes. Recent advancements in long-read sequencing and population sequencing have provided the opportunity to assemble sex-limited chromosomes without the traditional complicated experimental efforts.

FINDINGS

We introduce the first computational method, Sorting long Reads of Y or other sex-limited chromosome (SRY), which achieves improved assembly results compared to flow sorting. Specifically, SRY outperforms in the heterochromatic region and demonstrates comparable performance in other regions. Furthermore, SRY enhances the capabilities of the hybrid assembly software, resulting in improved continuity and accuracy.

CONCLUSIONS

Our method enables true complete genome assembly and facilitates downstream research of sex-limited chromosomes.

Adapting Nanopore Sequencing Basecalling Models for Modification Detection via Incremental Learning and Anomaly Detection

We leverage machine learning approaches to adapt nanopore sequencing basecallers for nucleotide modification detection. We first apply the incremental learning technique to improve the basecalling of modification-rich sequences, which are usually of high biological interests. With sequence backbones resolved, we further run anomaly detection on individual nucleotides to determine their modification status. By this means, our pipeline promises the single-molecule, single-nucleotide and sequence context-free detection of modifications. We benchmark the pipeline using control oligos, further apply it in the basecalling of densely-modified yeast tRNAs and E.coli genomic DNAs, the cross-species detection of N6-methyladenosine (m6A) in mammalian mRNAs, and the simultaneous detection of N1-methyladenosine (m1A) and m6A in human mRNAs. Our IL-AD workflow is available at: https://github.com/wangziyuan66/IL-AD.

Genome-wide chromatin interaction map for Trypanosoma cruzi

Trypanosomes are eukaryotic, unicellular parasites, such as Trypanosoma brucei, which causes sleeping sickness, and Trypanosoma cruzi, which causes Chagas disease. Genomes of these parasites comprise core regions and species-specific disruptive regions that encode multigene families of surface glycoproteins. Few transcriptional regulators have been identified in these parasites, and the role of spatial organization of the genome in gene expression is unclear. Here we mapped genome-wide chromatin interactions in T. cruzi using chromosome conformation capture (Hi-C), and we show that the core and disruptive regions form three-dimensional chromatin compartments named C and D. These chromatin compartments differ in levels of DNA methylation, nucleosome positioning and chromatin interactions, affecting genome expression dynamics. Our data reveal that the trypanosome genome is organized into chromatin-folding domains and transcription is affected by the local chromatin structure. We propose a model in which epigenetic mechanisms affect gene expression in trypanosomes.

Differential adenine methylation analysis reveals increased variability in 6mA in the absence of methyl-directed mismatch repair

IMPORTANCE

Methylation greatly influences the bacterial genome by guiding DNA repair and regulating pathogenic and stress-response phenotypes. But, the rate of epigenetic changes and their consequences on molecular phenotypes are underexplored. Through a detailed characterization of genome-wide adenine methylation in a commonly used laboratory strain of Escherichia coli, we reveal that mismatch repair deficient populations experience an increase in epimutations resulting in a genome-wide reduction of 6mA methylation in a manner consistent with genetic drift. Our findings highlight how methylation patterns evolve and the constraints on epigenetic evolution due to post-replicative DNA repair, contributing to a deeper understanding of bacterial genome evolution and how epimutations may introduce semi-permanent variation that can influence adaptation.

Genome-wide DNA methylation landscape of four Chinese populations and epigenetic variation linked to Tibetan high-altitude adaptation

DNA methylation (DNAm) is one of the major epigenetic mechanisms in humans and is important in diverse cellular processes. The variation of DNAm in the human population is related to both genetic and environmental factors. However, the DNAm profiles have not been investigated in the Chinese population of diverse ethnicities. Here, we performed double-strand bisulfite sequencing (DSBS) for 32 Chinese individuals representing four major ethnic groups including Han Chinese, Tibetan, Zhuang, and Mongolian. We identified a total of 604,649 SNPs and quantified DNAm at more than 14 million CpGs in the population. We found global DNAm-based epigenetic structure is different from the genetic structure of the population, and ethnic difference only partially explains the variation of DNAm. Surprisingly, non-ethnic-specific DNAm variations showed stronger correlation with the global genetic divergence than these ethnic-specific DNAm. Differentially methylated regions (DMRs) among these ethnic groups were found around genes in diverse biological processes. Especially, these DMR-genes between Tibetan and non-Tibetans were enriched around high-altitude genes including EPAS1 and EGLN1, suggesting DNAm alteration plays an important role in high-altitude adaptation. Our results provide the first batch of epigenetic maps for Chinese populations and the first evidence of the association of epigenetic changes with Tibetans' high-altitude adaptation.

Telomerase-independent survival leads to a mosaic of complex subtelomere rearrangements in Chlamydomonas reinhardtii

Telomeres and subtelomeres, the genomic regions located at chromosome extremities, are essential for genome stability in eukaryotes. In the absence of the canonical maintenance mechanism provided by telomerase, telomere shortening induces genome instability. The landscape of the ensuing genome rearrangements is not accessible by short-read sequencing. Here, we leverage Oxford Nanopore Technologies long-read sequencing to survey the extensive repertoire of genome rearrangements in telomerase mutants of the model green microalga Chlamydomonas reinhardtii In telomerase-mutant strains grown for hundreds of generations, most chromosome extremities were capped by short telomere sequences that were either recruited de novo from other loci or maintained in a telomerase-independent manner. Other extremities did not end with telomeres but only with repeated subtelomeric sequences. The subtelomeric elements, including rDNA, were massively rearranged and involved in breakage-fusion-bridge cycles, translocations, recombinations, and chromosome circularization. These events were established progressively over time and displayed heterogeneity at the subpopulation level. New telomere-capped extremities composed of sequences originating from more internal genomic regions were associated with high DNA methylation, suggesting that de novo heterochromatin formation contributes to the restoration of chromosome end stability in C. reinhardtii The diversity of alternative strategies present in the same organism to maintain chromosome integrity and the variety of rearrangements found in telomerase mutants are remarkable, and illustrate genome plasticity at short timescales.

BRCA genes as candidates for colorectal cancer genetic testing panel: systematic review and meta-analysis

BACKGROUND

Breast cancer susceptibility gene (BRCA) mutation carriers are at an increased risk for breast, ovarian, prostate and pancreatic cancers. However, the role of BRCA is unclear in colorectal cancer; the results regarding the association between BRCA gene mutations and colorectal cancer risk are inconsistent and even controversial. This study aimed to investigate whether BRCA1 and BRCA2 gene mutations are associated with colorectal cancer risk.

METHODS

In this systematic review, we searched PubMed/MEDLINE, Embase and Cochrane Library databases, adhering to PRISMA guidelines. Study quality was assessed using the Newcastle-Ottawa Scale (NOS). Unadjusted odds ratios (ORs) were used to estimate the probability of Breast Cancer Type 1 Susceptibility gene (BRCA1) and Breast Cancer Type 2 Susceptibility gene (BRCA2) mutations in colorectal cancer patients. The associations were evaluated using fixed effect models.

RESULTS

Fourteen studies were included in the systematic review. Twelve studies, including seven case-control and five cohort studies, were included in the meta-analysis. A significant increase in the frequency of BRCA1 and BRCA2 mutations was observed in patients with colorectal cancer [OR = 1.34, 95% confidence interval (CI) = 1.02-1.76, P = 0.04]. In subgroup analysis, colorectal cancer patients had an increased odds of BRCA1 (OR = 1.48, 95% CI = 1.10-2.01, P = 0.01) and BRCA2 (OR = 1.56, 95% CI = 1.06-2.30, P = 0.02) mutations.

CONCLUSIONS

BRCA genes are one of the genes that may increase the risk of developing colorectal cancer. Thus, BRCA genes could be potential candidates that may be included in the colorectal cancer genetic testing panel.

A tool for rapid, automated characterization of population epigenomics in plants

Epigenetic variation in plant populations is an important factor in determining phenotype and adaptation to the environment. However, while advances have been made in the molecular and computational methods to analyze the methylation status of a given sample of DNA, tools to profile and compare the methylomes of multiple individual plants or groups of plants at high resolution and low cost are lacking. Here, we describe a computational approach and R package (sounDMR) that leverages the benefits of long read nanopore sequencing to enable robust identification of differential methylation from complex experimental designs, as well as assess the variability within treatment groups and identify individual plants of interest. We demonstrate the utility of this approach by profiling a population of Arabidopsis thaliana exposed to a demethylating agent and identify genomic regions of high epigenetic variability between individuals. Given the low cost of nanopore sequencing devices and the ease of sample preparation, these results show that high resolution epigenetic profiling of plant populations can be made more broadly accessible in plant breeding and biotechnology.

The Global Changes of N6-methyldeoxyadenosine in Response to Low Temperature in Arabidopsis thaliana and Rice

N6-methyldeoxyadenosine (6mA) is a recently discovered DNA modification involved in regulating plant adaptation to abiotic stresses. However, the mechanisms and changes of 6mA under cold stress in plants are not yet fully understood. Here, we conducted a genome-wide analysis of 6mA and observed that 6mA peaks were predominantly present within the gene body regions under both normal and cold conditions. In addition, the global level of 6mA increased both in Arabidopsis and rice after the cold treatment. The genes that exhibited an up-methylation showed enrichment in various biological processes, whereas there was no significant enrichment observed among the down-methylated genes. The association analysis revealed a positive correlation between the 6mA level and the gene expression level. Joint analysis of the 6mA methylome and transcriptome of Arabidopsis and rice unraveled that fluctuations in 6mA levels caused by cold exposure were not correlated to changes in transcript levels. Furthermore, we discovered that orthologous genes modified by 6mA showed high expression levels; however, only a minor amount of differentially 6mA-methylated orthologous genes were shared between Arabidopsis and rice under low-temperature conditions. In conclusion, our study provides information on the role of 6mA in response to cold stress and reveals its potential for regulating the expression of stress-related genes.

Applications of long-read sequencing to Mendelian genetics

Advances in clinical genetic testing, including the introduction of exome sequencing, have uncovered the molecular etiology for many rare and previously unsolved genetic disorders, yet more than half of individuals with a suspected genetic disorder remain unsolved after complete clinical evaluation. A precise genetic diagnosis may guide clinical treatment plans, allow families to make informed care decisions, and permit individuals to participate in N-of-1 trials; thus, there is high interest in developing new tools and techniques to increase the solve rate. Long-read sequencing (LRS) is a promising technology for both increasing the solve rate and decreasing the amount of time required to make a precise genetic diagnosis. Here, we summarize current LRS technologies, give examples of how they have been used to evaluate complex genetic variation and identify missing variants, and discuss future clinical applications of LRS. As costs continue to decrease, LRS will find additional utility in the clinical space fundamentally changing how pathological variants are discovered and eventually acting as a single-data source that can be interrogated multiple times for clinical service.

Navigating the pitfalls of mapping DNA and RNA modifications

Chemical modifications to nucleic acids occur across the kingdoms of life and carry important regulatory information. Reliable high-resolution mapping of these modifications is the foundation of functional and mechanistic studies, and recent methodological advances based on next-generation sequencing and long-read sequencing platforms are critical to achieving this aim. However, mapping technologies may have limitations that sometimes lead to inconsistent results. Some of these limitations are technical in nature and specific to certain types of technology. Here, however, we focus on common (yet not always widely recognized) pitfalls that are shared among frequently used mapping technologies and discuss strategies to help technology developers and users mitigate their effects. Although the emphasis is primarily on DNA modifications, RNA modifications are also discussed.

Exploring the crop epigenome: a comparison of DNA methylation profiling techniques

Epigenetic modifications play a vital role in the preservation of genome integrity and in the regulation of gene expression. DNA methylation, one of the key mechanisms of epigenetic control, impacts growth, development, stress response and adaptability of all organisms, including plants. The detection of DNA methylation marks is crucial for understanding the mechanisms underlying these processes and for developing strategies to improve productivity and stress resistance of crop plants. There are different methods for detecting plant DNA methylation, such as bisulfite sequencing, methylation-sensitive amplified polymorphism, genome-wide DNA methylation analysis, methylated DNA immunoprecipitation sequencing, reduced representation bisulfite sequencing, MS and immuno-based techniques. These profiling approaches vary in many aspects, including DNA input, resolution, genomic region coverage, and bioinformatics analysis. Selecting an appropriate methylation screening approach requires an understanding of all these techniques. This review provides an overview of DNA methylation profiling methods in crop plants, along with comparisons of the efficacy of these techniques between model and crop plants. The strengths and limitations of each methodological approach are outlined, and the importance of considering both technical and biological factors are highlighted. Additionally, methods for modulating DNA methylation in model and crop species are presented. Overall, this review will assist scientists in making informed decisions when selecting an appropriate DNA methylation profiling method.

Integrated fragmentomic profile and 5-Hydroxymethylcytosine of capture-based low-pass sequencing data enables pan-cancer detection via cfDNA

BACKGROUND

Using epigenetic markers and fragmentomics of cell-free DNA for cancer detection has been proven applicable.

METHODS

We further investigated the diagnostic potential of combining two features (epigenetic markers and fragmentomic information) of cell-free DNA for detecting various types of cancers. To do this, we extracted cfDNA fragmentomic features from 191 whole-genome sequencing data and studied them in 396 low-pass 5hmC sequencing data, which included four common cancer types and control samples.

RESULTS

In our analysis of 5hmC sequencing data from cancer samples, we observed aberrant ultra-long fragments (220-500 bp) that differed from normal samples in terms of both size and coverage profile. These fragments played a significant role in predicting cancer. Leveraging the ability to detect cfDNA hydroxymethylation and fragmentomic markers simultaneously in low-pass 5hmC sequencing data, we developed an integrated model that incorporated 63 features representing both fragmentomic features and hydroxymethylation signatures. This model achieved high sensitivity and specificity for pan-cancer detection (88.52% and 82.35%, respectively).

CONCLUSION

We showed that fragmentomic information in 5hmC sequencing data is an ideal marker for cancer detection and that it shows high performance in low-pass sequencing data.

6mA DNA Methylation on Genes in Plants Is Associated with Gene Complexity, Expression and Duplication

N6-methyladenine (6mA) DNA methylation has emerged as an important epigenetic modification in eukaryotes. Nevertheless, the evolution of the 6mA methylation of homologous genes after species and after gene duplications remains unclear in plants. To understand the evolution of 6mA methylation, we detected the genome-wide 6mA methylation patterns of four lotus plants (Nelumbo nucifera) from different geographic origins by nanopore sequencing and compared them to patterns in Arabidopsis and rice. Within lotus, the genomic distributions of 6mA sites are different from the widely studied 5mC methylation sites. Consistently, in lotus, Arabidopsis and rice, 6mA sites are enriched around transcriptional start sites, positively correlated with gene expression levels, and preferentially retained in highly and broadly expressed orthologs with longer gene lengths and more exons. Among different duplicate genes, 6mA methylation is significantly more enriched and conserved in whole-genome duplicates than in local duplicates. Overall, our study reveals the convergent patterns of 6mA methylation evolution based on both lineage and duplicate gene divergence, which underpin their potential role in gene regulatory evolution in plants.

The Chlamydomonas Genome Project, version 6: Reference assemblies for mating-type plus and minus strains reveal extensive structural mutation in the laboratory

Five versions of the Chlamydomonas reinhardtii reference genome have been produced over the last two decades. Here we present version 6, bringing significant advances in assembly quality and structural annotations. PacBio-based chromosome-level assemblies for two laboratory strains, CC-503 and CC-4532, provide resources for the plus and minus mating-type alleles. We corrected major misassemblies in previous versions and validated our assemblies via linkage analyses. Contiguity increased over ten-fold and >80% of filled gaps are within genes. We used Iso-Seq and deep RNA-seq datasets to improve structural annotations, and updated gene symbols and textual annotation of functionally characterized genes via extensive manual curation. We discovered that the cell wall-less classical reference strain CC-503 exhibits genomic instability potentially caused by deletion of the helicase RECQ3, with major structural mutations identified that affect >100 genes. We therefore present the CC-4532 assembly as the primary reference, although this strain also carries unique structural mutations and is experiencing rapid proliferation of a Gypsy retrotransposon. We expect all laboratory strains to harbor gene-disrupting mutations, which should be considered when interpreting and comparing experimental results. Collectively, the resources presented here herald a new era of Chlamydomonas genomics and will provide the foundation for continued research in this important reference organism.

Detection of DNA Modification Using Nanopore Sequencers

DNA modification is a crucial factor of epigenetic modification and has vital functions for gene regulation and phenotype control. A profound understanding of DNA modification requires precise mapping of the modified bases on genomic DNA. In addition to methods such as bisulfite sequencing and single-molecule real-time (SMRT) sequencing of PacBio sequencers, nanopore sequencers can be also utilized for the detection of DNA modification. Here, I will briefly review the three methods for the detection of DNA modification with nanopore sequencers and introduce a protocol using MinION and Megalodon.

Recent developments in visible light induced polymerization towards its application to nanopores

Visible light induced polymerizations are a strongly emerging field in recent years. Besides the often mild reaction conditions, visible light offers advantages of spatial and temporal control over chain growth, which makes visible light ideal for functionalization of surfaces and more specifically of nanoscale pores. Current challenges in nanopore functionalization include, in particular, local and highly controlled polymer functionalizations. Using spatially limited light sources such as lasers or near field modes for light-induced polymer functionalization is envisioned to allow local functionalization of nanopores and thereby improve nanoporous material performance. These light sources are usually providing visible light while classical photopolymerizations are mostly based on UV-irradiation. In this review, we highlight developments in visible light induced polymerizations and especially in visible light induced controlled polymerizations as well as their potential for nanopore functionalization. Existing examples of visible light induced polymerizations in nanopores are emphasized.

Studying Whole-Genome Duplication Using Experimental Evolution of Chlamydomonas

In this chapter, we present the use of Chlamydomonas reinhardtii in experiments designed to study the evolutionary impacts of whole genome duplication. We shortly introduce the algal species and depict why it is an excellent model for experimental evolution. Subsequently, we discuss the most relevant steps and methods in the design of a ploidy-related Chlamydomonas experiment. These steps include strain selection, ploidy determination, different methods of making diplo- and polyploid Chlamydomonas cells, replication, culturing conditions, preservation, and the ways to quantify phenotypic and genotypic change.

Solid-State Nanopore Array: Manufacturing and Applications

Nanopore brings extraordinary properties for a variety of potential applications in various industrial sectors. Since manufacturing of solid-state nanopore is first reported in 2001, solid-state nanopore has become a hot topic in the recent years. An increasing number of manufacturing methods have been reported, with continuously decreased sizes from hundreds of nanometers at the beginning to ≈1 nm until recently. To enable more robust, sensitive, and reliable devices required by the industry, researchers have started to explore the possible methods to manufacture nanopore array which presents unprecedented challenges on the fabrication efficiency, accuracy and repeatability, applicable materials, and cost. As a result, the exploration of fabrication of nanopore array is still in the fledging period with various bottlenecks. In this article, a wide range of methods of manufacturing nanopores are summarized along with their achievable morphologies, sizes, inner structures for characterizing the main features, based on which the manufacturing of nanopore array is further addressed. To give a more specific idea on the potential applications of nanopore array, some representative practices are introduced such as DNA/RNA sequencing, energy conversion and storage, water desalination, nanosensors, nanoreactors, and dialysis.

Phasing DNA Methylation

Haplotyping enables the study of allele-specific events. Heterozygous variants, primarily single nucleotide variants (SNVs), enable the assignment of the paternal and maternal origin of the chromosomes and are widely employed to phase sequencing reads to their haplotype of origin. Certain long-read technologies enable the detection of both the DNA sequence and DNA modifications. These long reads and their inherent methylation information are suitable for genome-wide haplotyping and allele-specific DNA methylation analysis. Here, we describe the workflow to phase reads and DNA methylation using nanopore sequencing.

CNN6mA: Interpretable neural network model based on position-specific CNN and cross-interactive network for 6mA site prediction

N6-methyladenine (6mA) plays a critical role in various epigenetic processing including DNA replication, DNA repair, silencing, transcription, and diseases such as cancer. To understand such epigenetic mechanisms, 6 mA has been detected by high-throughput technologies on a genome-wide scale at single-base resolution, together with conventional methods such as immunoprecipitation, mass spectrometry and capillary electrophoresis, but these experimental approaches are time-consuming and laborious. To complement these problems, we have developed a CNN-based 6 mA site predictor, named CNN6mA, which proposed two new architectures: a position-specific 1-D convolutional layer and a cross-interactive network. In the position-specific 1-D convolutional layer, position-specific filters with different window sizes were applied to an inquiry sequence instead of sharing the same filters over all positions in order to extract the position-specific features at different levels. The cross-interactive network explored the relationships between all the nucleotide patterns within the inquiry sequence. Consequently, CNN6mA outperformed the existing state-of-the-art models in many species and created the contribution score vector that intelligibly interpret the prediction mechanism. The source codes and web application in CNN6mA are freely accessible at https://github.com/kuratahiroyuki/CNN6mA.git and http://kurata35.bio.kyutech.ac.jp/CNN6mA/, respectively.

Simultaneous profiling of histone modifications and DNA methylation via nanopore sequencing

The interplay between histone modifications and DNA methylation drives the establishment and maintenance of the cellular epigenomic landscape, but it remains challenging to investigate the complex relationship between these epigenetic marks across the genome. Here we describe a nanopore-sequencing-based-method, nanoHiMe-seq, for interrogating the genome-wide localization of histone modifications and DNA methylation from single DNA molecules. nanoHiMe-seq leverages a nonspecific methyltransferase to exogenously label adenine bases proximal to antibody-targeted modified nucleosomes in situ. The labelled adenines and the endogenous methylated CpG sites are simultaneously detected on individual nanopore reads using a hidden Markov model, which is implemented in the nanoHiMe software package. We demonstrate the utility, robustness and sensitivity of nanoHiMe-seq by jointly profiling DNA methylation and histone modifications at low coverage depths, concurrently determining phased patterns of DNA methylation and histone modifications, and probing the intrinsic connectivity between these epigenetic marks across the genome.

Single-Cell DNA Methylation Analysis in Cancer

Morphological, transcriptomic, and genomic defects are well-explored parameters of cancer biology. In more recent years, the impact of epigenetic influences, such as DNA methylation, is becoming more appreciated. Aberrant DNA methylation has been implicated in many types of cancers, influencing cell type, state, transcriptional regulation, and genomic stability to name a few. Traditionally, large populations of cells from the tissue of interest are coalesced for analysis, producing averaged methylome data. Considering the inherent heterogeneity of cancer, analysing populations of cells as a whole denies the ability to discover novel aberrant methylation patterns, identify subpopulations, and trace cell lineages. Due to recent advancements in technology, it is now possible to obtain methylome data from single cells. This has both research and clinical implications, ranging from the identification of biomarkers to improved diagnostic tools. As with all emerging technologies, distinct experimental, bioinformatic, and practical challenges present themselves. This review begins with exploring the potential impact of single-cell sequencing on understanding cancer biology and how it could eventually benefit a clinical setting. Following this, the techniques and experimental approaches which made this technology possible are explored. Finally, the present challenges currently associated with single-cell DNA methylation sequencing are described.

Active learning for efficient analysis of high-throughput nanopore data

MOTIVATION

As the third-generation sequencing technology, nanopore sequencing has been used for high-throughput sequencing of DNA, RNA, and even proteins. Recently, many studies have begun to use machine learning technology to analyze the enormous data generated by nanopores. Unfortunately, the success of this technology is due to the extensive labeled data, which often suffer from enormous labor costs. Therefore, there is an urgent need for a novel technology that can not only rapidly analyze nanopore data with high-throughput, but also significantly reduce the cost of labeling. To achieve the above goals, we introduce active learning to alleviate the enormous labor costs by selecting the samples that need to be labeled. This work applies several advanced active learning technologies to the nanopore data, including the RNA classification dataset (RNA-CD) and the Oxford Nanopore Technologies barcode dataset (ONT-BD). Due to the complexity of the nanopore data (with noise sequence), the bias constraint is introduced to improve the sample selection strategy in active learning. Results: The experimental results show that for the same performance metric, 50% labeling amount can achieve the best baseline performance for ONT-BD, while only 15% labeling amount can achieve the best baseline performance for RNA-CD. Crucially, the experiments show that active learning technology can assist experts in labeling samples, and significantly reduce the labeling cost. Active learning can greatly reduce the dilemma of difficult labeling of high-capacity nanopore data. We hope active learning can be applied to other problems in nanopore sequence analysis.

AVAILABILITY AND IMPLEMENTATION

The main program is available at https://github.com/guanxiaoyu11/AL-for-nanopore.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Plant synthetic epigenomic engineering for crop improvement

Efforts have been directed to redesign crops with increased yield, stress adaptability, and nutritional value through synthetic biology-the application of engineering principles to biology. A recent expansion in our understanding of how epigenetic mechanisms regulate plant development and stress responses has unveiled a new set of resources that can be harnessed to develop improved crops, thus heralding the promise of "synthetic epigenetics." In this review, we summarize the latest advances in epigenetic regulation and highlight how innovative sequencing techniques, epigenetic editing, and deep learning-driven predictive tools can rapidly extend these insights. We also proposed the future directions of synthetic epigenetics for the development of engineered smart crops that can actively monitor and respond to internal and external cues throughout their life cycles.

Multi-modality artificial intelligence in digital pathology

In common medical procedures, the time-consuming and expensive nature of obtaining test results plagues doctors and patients. Digital pathology research allows using computational technologies to manage data, presenting an opportunity to improve the efficiency of diagnosis and treatment. Artificial intelligence (AI) has a great advantage in the data analytics phase. Extensive research has shown that AI algorithms can produce more up-to-date and standardized conclusions for whole slide images. In conjunction with the development of high-throughput sequencing technologies, algorithms can integrate and analyze data from multiple modalities to explore the correspondence between morphological features and gene expression. This review investigates using the most popular image data, hematoxylin-eosin stained tissue slide images, to find a strategic solution for the imbalance of healthcare resources. The article focuses on the role that the development of deep learning technology has in assisting doctors' work and discusses the opportunities and challenges of AI.

Nanopore Detection Assisted DNA Information Processing

The deoxyribonucleotide (DNA) molecule is a stable carrier for large amounts of genetic information and provides an ideal storage medium for next-generation information processing technologies. Technologies that process DNA information, representing a cross-disciplinary integration of biology and computer techniques, have become attractive substitutes for technologies that process electronic information alone. The detailed applications of DNA technologies can be divided into three components: storage, computing, and self-assembly. The quality of DNA information processing relies on the accuracy of DNA reading. Nanopore detection allows researchers to accurately sequence nucleotides and is thus widely used to read DNA. In this paper, we introduce the principles and development history of nanopore detection and conduct a systematic review of recent developments and specific applications in DNA information processing involving nanopore detection and nanopore-based storage. We also discuss the potential of artificial intelligence in nanopore detection and DNA information processing. This work not only provides new avenues for future nanopore detection development, but also offers a foundation for the construction of more advanced DNA information processing technologies.

Rapid and simple analysis of short and long sequencing reads using DuesselporeTM

Transcriptome analysis experiments enable researchers to gain extensive insights into the molecular mechanisms underlying cell physiology and disease. Oxford Nanopore Technologies (ONT) has recently been developed as a fast, miniaturized, portable, and cost-effective alternative to next-generation sequencing (NGS). However, RNA-Seq data analysis software that exploits ONT portability and allows scientists to easily analyze ONT data everywhere without bioinformatics expertise is not widely available. We developed Duesselpore^TM, an easy-to-follow deep sequencing workflow that runs as a local webserver and allows the analysis of ONT data everywhere without requiring additional bioinformatics tools or internet connection. Duesselpore^TM output includes differentially expressed genes and further downstream analyses, such as variance heatmap, disease and gene ontology plots, gene concept network plots, and exports customized pathways for different cellular processes. We validated Duesselpore^TM by analyzing the transcriptomic changes induced by PCB126, a dioxin-like PCB, and a potent aryl hydrocarbon receptor (AhR) agonist in human HaCaT keratinocytes, a well-characterized model system. Duesselpore^TM was specifically developed to analyze ONT data, but we also implemented NGS data analysis. Duesselpore^TM is compatible with Linux, Microsoft, and Mac operating systems and allows convenient, reliable, and cost-effective analysis of ONT and NGS data.

Assessing comparative importance of DNA sequence and epigenetic modifications on gene expression using a deep convolutional neural network

Gene expression is regulated at both transcriptional and post-transcriptional levels. DNA sequence and epigenetic modifications are key factors which regulate gene transcription. Understanding their complex interactions and their respective contributions to gene expression regulation remains a challenge in biological studies. We have developed iSEGnet, a framework of deep convolutional neural network to predict mRNA abundance using the information on DNA sequences as well as epigenetic modifications within genes and their cis-regulatory regions. We demonstrate that our framework outperforms other machine learning models in terms of predicting mRNA abundance using transcriptional and epigenetic profiles from six distinct cell lines/types chosen from the ENCODE. The analysis from the learned models also reveals that specific regions around promotors and transcription termination sites are most important for gene expression regulation. Using the method of Integrated Gradients, we identify narrow segments in these regions which are most likely to impact gene expression for a specific epigenetic modification. We further show that these identified segments are enriched in known active regulatory regions by comparing the transcription factor binding sites obtained via ChIP-seq. Moreover, we demonstrate how iSEGnet can uncover potential transcription factors that have regulatory functions in cancer using two cancer multi-omics data.

Machine Learning for Electrocatalyst and Photocatalyst Design and Discovery

Electrocatalysts and photocatalysts are key to a sustainable future, generating clean fuels, reducing the impact of global warming, and providing solutions to environmental pollution. Improved processes for catalyst design and a better understanding of electro/photocatalytic processes are essential for improving catalyst effectiveness. Recent advances in data science and artificial intelligence have great potential to accelerate electrocatalysis and photocatalysis research, particularly the rapid exploration of large materials chemistry spaces through machine learning. Here a comprehensive introduction to, and critical review of, machine learning techniques used in electrocatalysis and photocatalysis research are provided. Sources of electro/photocatalyst data and current approaches to representing these materials by mathematical features are described, the most commonly used machine learning methods summarized, and the quality and utility of electro/photocatalyst models evaluated. Illustrations of how machine learning models are applied to novel electro/photocatalyst discovery and used to elucidate electrocatalytic or photocatalytic reaction mechanisms are provided. The review offers a guide for materials scientists on the selection of machine learning methods for electrocatalysis and photocatalysis research. The application of machine learning to catalysis science represents a paradigm shift in the way advanced, next-generation catalysts will be designed and synthesized.