Translational Opportunities for Microfluidic Technologies to Enable Precision Epigenomics

Abstract

Personalizing health care by taking genetic, environmental and lifestyle factors into account is central to modern medicine. The crucial and pervasive roles epigenetic factors play in shaping gene–environment interactions are now well recognized. However, identifying robust epigenetic biomarkers and translating them to clinical tests has been difficult due in part to limitations of available platforms to detect epigenetic features genome-wide (epigenomic assays). This Feature introduces several important prospects for precision epigenomics, highlights capabilities and limitations of current laboratory technologies, and emphasizes opportunities for microfluidic tools to facilitate translation of epigenetic analyses to the clinic, with a particular focus on methods to profile gene-associated histone modifications and their impacts on chromatin structure and gene expression.

Graphical Abstract

I. EPIGENOMIC IMPLICATIONS FOR PRECISION MEDICINE

Precision medicine is a modern medicine approach that considers a patient’s biochemical markers indicative of disease risk or progression in order to design an optimized prevention and treatment strategy tailored for a specific individual.^{1, 2} As a critical innovation for precision medicine, next-generation sequencing (NGS) has led to numerous breakthroughs in terms of revealing specific tumor-driving mutations and assessing patient heterogeneity.³ Beyond the linear sequence of DNA, environmental factors may also influence phenotype.⁴ Perhaps the most striking example of non-genetic mechanisms shaping phenotypic diversity is the divergence of the transcriptome and other phenotypic traits in monozygotic twins occurring likely in response to environmental cues.^{5, 6} Sensitive to environmental insults, a complex regulatory network of “epigenetic” factors acts beyond the genome and can even cause diseases such as cancer, especially when dysregulated.^{7, 8} Thus, elucidating details of the epigenome and their potential heterogeneity in disease development and progression is emerging as an exciting new aspect of precision medicine.

Understanding the Epigenome.

The epigenome encompasses multiple layers of interacting regulatory elements that define phenotypic variation beyond what the DNA sequence encodes.^{9, 10} Upon establishment, epigenetic features impact cellular differentiation and development, and can persist through multiple rounds of mitotic or even meiotic divisions.^11–13 However stable, these epigenetic traits are actually reversible through germline reprogramming and after fertilization as totipotent stem cells are formed.^{13, 14} Such reversibility has enabled the generation of induced pluripotent stem cells.¹⁵ Beyond differentiation and development, these epigenetic factors play a major role in transcriptional regulation in disease pathways, further underscoring the importance of understanding their molecular bases.¹⁶

Understanding chromatin structure is crucial to understanding the epigenome, as chromatin is the template of epigenetic regulation (Fig. 1). The most fundamental unit of chromatin is the nucleosome. Each nucleosome has a core of octameric protein complexes that contain two copies of four histone proteins (H2A, H2B, H3, and H4) and is surrounded by ~147 base pairs (1.67 turns) of DNA.¹⁷ With linker DNA connecting the nucleosomes, chromatin compaction and decompaction are achieved by higher order assembly and disassembly of nucleosomes that effectively control accessibility of particular genomic regions to transcriptional machinery (heterochromatin and euchromatin).¹⁸ Epigenetic modifications provide a genome-wide indexing system that, through the actions of chromatin remodelers and chaperones, regulates the use of the DNA template for transcription, replication, and repair, and can be indicative of the functional state of chromatin during analysis.^{11, 18}

Figure 1.

Overview of the chromatin and related common epigenetic marks. Histone H2A, H2B, H3 and H4 are represented by the same two colors, alternating to reveal the octamer structure.

The best understood epigenomic factor is DNA methylation. Cytosine is frequently covalently modified to 5-methyl cytosine (5mC), particularly in a CG dinucleotide context. Oxidative derivatives of 5mC, including 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC) and 5-carboxylcytosine (5caC), although less stable and less abundant, also have functional significance besides representing demethylation intermediates.¹⁹ 5mC at CpG islands (DNA regions with a relatively high frequency of CG dinucleotide sites) near gene promoters interferes with the binding of transcription machinery, repressing gene expression.²⁰ 5hmC regions most commonly associate with poised and active enhancers, actively transcribed gene bodies, and “bivalent” promoters bearing both activating and repressive histone marks, which are characteristic of developmentally regulated genes poised for expression or frank repression (Fig. 1, only showing 5hmC at active enhancers and promoters).²¹ DNA methyltransferases (DNMTs) regulate the installation and maintenance of the methyl groups, in conjunction with Ten-Eleven Translocation (TET) proteins and members of the Base Excision Repair (BER) pathways, which further oxidize and ultimately erase the methylation marks.²⁰

Variation of histone proteins is another major epigenomic mechanism regulating gene expression.¹² Genetically encoded variants of histone subunits such as H3.3 play critical roles in germline cell maturation and other processes by aiding in the recruitment of transcription factors and chromatin remodeling enzymes with downstream effects.¹² Similarly, post-translational modifications (PTMs) to histones directly impact their binding interactions with DNA and other biomolecules. For example, acetylation at histone tails decreases the positive charge of the histone, thus reducing the magnitude of its electrostatic attraction to DNA and assisting in increasing DNA accessibility. Methylation at lysine and arginine residues may act as activating or repressive marks by changing the chromatin structure and, as a result, changing the binding properties of transcription factors and chaperones.^{18, 22} Dynamically regulated by histone acetyltransferases, deacetylases and methyltransferases (HATs, HDACs, and HMTs), histone modifications and variants (the “histone code”) synergistically interact with other epigenetic mechanisms to regulate DNA transcription, replication, and repair.^{18, 23}

Additionally, non-coding RNAs (ncRNAs) may also regulate gene expression. Some small RNAs such as Piwi-interacting RNAs (piRNAs) can recruit chromatin remodelers and other enzymes to modify histones.^{18, 24} Others like microRNAs (miRNAs, 17–25 nt) directly affect gene expression by binding to messenger RNA (mRNA) in conjunction with the RNA-induced silencing complex (RISC) to degrade mRNAs and repress translation.^{25, 26} Another class of RNA termed long non-coding RNA (lncRNA, >200 nt) can sequester regulatory small RNAs or directly participate in chromatin remodeling.^{25, 27} Active enhancers are transcribed into a class of lncRNAs (eRNAs), which regulate gene expression by altering the physicochemical properties of enhancer-promoter units.²⁸ Furthermore, chemical modifications to RNA have also been increasingly demonstrated as a mechanism for epigenetic regulation.²⁹

The combined action of DNA modifications, histone variants, histone modifications, and non-coding RNAs facilitate a complex regulatory network controlling gene expression by dynamically altering the three-dimensional conformation and accessibility of chromatin in a gene-specific manner. This three-dimensional architecture also affects DNA replication and repair, and overall, holds significant consequences for heath and disease.^{30, 31}

Epigenetics and Disease.

Because of their involvement in gene regulation and DNA replication/repair, dysregulation or aberrant function of epigenetic regulators may lead to detrimental consequences including carcinogenesis.^{22, 32} Epigenetic repression of tumor suppressors often arises from promoter hypermethylation of normally unmethylated CpG islands.³³ Likewise, dysregulation of the H19 lncRNA can downregulate the tumor suppressor gene retinoblastoma in human colorectal cancers.³⁴ Beyond cancer, epigenetic dysregulation can lead to many other diseases including obesity, autoimmune diseases, and neurological disorders.³⁵

Mirroring their multifaceted roles in pathogenesis, epigenetic factors also present promising therapeutic opportunities. Currently, most epigenetic-based therapeutic strategies focus on two major approaches: reprogramming aberrant epigenomic regulation to reestablish homeostasis, or utilizing epigenetic pathways to directly induce apoptosis or increase drug sensitivity for destruction of aberrant cells.³⁶ For example, weak HDAC inhibitors promote repression and reversion of the glycolytic phenotype in cancer cells, thus repressing tumor growth and invasion.³² On the other hand, several FDA-approved HDAC inhibitors (e.g., belinostat and vorinostat) selectively disrupt oncogenic pathways to trigger apoptosis in cancer cells. Furthermore, epigenetic factors may find utility as critical biomarkers to enhance diagnosis, offer predictive value to guide therapies, and monitor response to treatment. For instance, a recent study successfully demonstrated a predictive association between DNA methylation profile and potential for relapse in children with acute myeloid leukemia.³⁷

As efforts continue to discover epigenetic mechanisms involved in disease development, the potential for epigenetically driven precision therapies becomes increasingly within reach. The following sections describe the current common technologies for profiling epigenomic signatures, mainly based on NGS for its capability to achieve genome-wide profiling, and the gaps between promise and realization of precision epigenomics. PCR-based methods such as methylation-specific PCR (MSP) for DNA methylation detection in targeted genomic regions are not discussed here, but they have been widely used before NGS-based methods gained its popularity and readers are directed to the following excellent references describing these methods.^{38, 39}

II. CURRENT TECHNOLOGIES FOR EPIGENOMIC PROFILING

Direct Sequencing to Investigate ncRNAs.

RNA sequencing (RNA-seq) has been adopted to reveal the roles ncRNAs play in phenotypic regulation.⁴⁰ The general workflow starts with reverse transcribing a collection of RNA samples to cDNA libraries followed by adapter ligation and optional amplification by polymerase chain reaction (PCR). After sequencing the libraries and mapping the reads, quantifying read numbers mapped to a specific gene enables the assessment of the level of gene expression.⁴¹ Concentrating targets by size selection out of a total RNA sample or deconvolving the total reads generally enables direct profiling of small ncRNAs, whereas lncRNAs may need an additional step of sample fragmentation to ensure size compatibility with deep-sequencing technologies.^{42, 43} Although it may suffer from potential bias during library preparation and bioinformatic challenges, RNA-seq offers several analytical advantages including single base resolution, low background signals and large dynamic ranges, as well as the capability of discovering novel ncRNAs.^{41, 42}

Structure Conversion-based Methods Followed by Sequencing to Detect DNA Modifications.

The prevalent methods to profile DNA methylation are a collection of bisulfite sequencing strategies where the structures of modified cytosines are chemically altered. Whole-genome bisulfite sequencing (WGBS) treats DNA with a bisulfite reagent to convert regular cytosine, 5fC, and 5caC to uracil, while leaving 5mC and 5hmC intact to be detected as cytosines. Subsequent sequencing profiles genome-wide 5mC and 5hmC down to single nucleotide resolution.^{44, 45} Reduced representation bisulfite sequencing (RRBS) reduces cost and improves efficiency over WGBS by concentrating on CpG-rich regions via restriction enzyme digestion (e.g., with MspI and BglII) prior to bisulfite treatment.^{46, 47} Thus RRBS enriches CpG-rich regions and as such does not as extensively cover enhancers and intronic regions as WGBS.¹⁹ Additionally, Tet-assisted bisulfite sequencing (TAB-seq) specifically detects 5hmC at single-base resolution by protecting 5hmC with glycosylation and oxidizing other cytosine residues with Tet enzyme. Therefore, only 5hmC remains intact after bisulfite treatment.^{48, 49} Likewise, oxidative bisulfite sequencing (oxBS-Seq) differentiates 5mC and 5hmC by specifically oxidizing 5hmC before bisulfite treatment to obtain a positive readout of pure 5mC. Strategies specifically detecting 5fC, 5caC, and other DNA modifications have also emerged.¹⁹ Since their initial demonstration, applications of these methods have extended our understandings of the biological roles of DNA modifications.

Affinity-Based Methods Followed by Sequencing to Detect DNA and Histone Modifications and DNA-binding Proteins.

In addition to methods based on structure conversion, epigenetic studies have widely exploited affinity-based methods to enrich DNA modifications for underlying biological impact elucidation.⁵⁰ Methylated DNA immunoprecipitation (MeDIP) enriches 5mC DNA regions with 5mC-specific antibodies prior to sequencing to generate high-resolution whole-genome DNA methylation profiles (MeDIP-seq; Fig. 2).^{51, 52} Methods using the MBD of MBD protein 2 (MBD2) to precipitate methylated DNA regions can also identify DNA methylation patterns when integrated with sequencing (i.e. MBD-seq based methods).^{53, 54} Compared to WGBS, TAB-seq, and oxBS-seq, affinity-based methods have biased coverage towards CpG-rich regions; however, MeDIP- and MBD-seq provide information on these regions with lower costs and higher efficiency.^{55, 56}

Figure 2.

Workflows of selected affinity-based methods for assessing genomic loci associated with specific chromatin modifications.

Furthermore, affinity-based immunoprecipitation also helps profile histone modifications and other DNA-binding proteins. The gold standard to identify such protein-DNA interactions is chromatin immunoprecipitation followed by sequencing (ChIP-seq; Fig.2). Using antibodies targeting specific histone modifications or DNA-binding proteins of interest, fragmented chromatin carrying the target will be captured and have the associated DNA purified for library preparation and sequencing.^57–59 Comparing the sequencing results of the ChIP library to an input control allows the assessment of enrichment specificity and efficiency. Subsequently, variations of ChIP-seq have emerged and brought improvements. These include ChIP-exo for mapping resolution enhancement⁶⁰ and ChIP-Rx for quantitative comparison among multiple ChIP-seq runs.⁶¹ The protocol called “Cleavage Under Targets and Release Using Nuclease” (CUT&RUN) enables low-background and base pair-resolution profiling of chromatin modification and protein binding sites genome-wide with low inputs of cells.⁶² This protocol utilizes a fusion between protein A and Micrococcal nuclease (MNase) to recognize antibodies bound to targeted epigenetic factors, and the spatially-confined nuclease activity selectively releases only the regions of chromatin involved in the targeted interaction. Replacing MNase with Tn5 transposase in the fusion protein (Cleavage Under Targets and Tagmentation; CUT&Tag) enables simultaneous DNA fragmentation and adaptor tagging, further simplifying the workflow and reducing input requirements.⁶³

Enzymatic and Chemical Processing Methods Followed by Sequencing to Assess Chromatin Accessibility and Nucleosome Position.

Various methods have been developed to study nucleosome and chromatin dynamics in epigenetics. MNase is an endo-exonuclease that digests chromatin with minimal sequence preference into (mono-)nucleosomes by cleaving exposed DNA and DNA ends until encountering a barrier (e.g., nucleosome).^{64, 65} Combining MNase digestion with high-throughput sequencing (MNase-seq) enables localization of DNA-binding proteins and nucleosome positioning at single base pair resolution (Fig. 3).⁶⁶ Moreover, the assay for transposase-accessible chromatin using sequencing (ATAC-seq) utilizes hyperactive Tn5 transposase to fragment and tag the genome with sequencing adapters simultaneously (Fig. 3).^67,68 ATAC-seq enables fast genome-wide mapping of active regulatory elements, nucleosome positioning and chromatin accessibility at the same time. This method has largely replaced other traditional open chromatin assays (e.g., DNase-seq and FAIRE-seq). MNase-seq and ATAC-seq have enhanced the community’s understanding of the regulation of nucleosome distribution and chromatin accessibility, as well as their roles in a variety of biological processes.

Figure 3.

Workflows of selected methods to analyze nucleosome distribution and open chromatin.

Hybridization-Based Technologies.

With the rapid advancement of next-generation sequencing, the conventional hybridization-based microarrays have become less popular. However, Illumina’s Infinium Methylation EPIC Kit (EPIC array) is one major remaining microarray method used in epigenetic studies for genome-wide CpG methylation detection.⁶⁹ In this assay, bisulfite converted DNA hybridizes with the carefully designed probes that target the sequences of interest but are different only at the 3’ end per tested CpG site, one corresponding to unmethylated C and the other methylated. Subsequent fluorescence staining and detection records and calculates the fluorescent intensity ratios between these two types of probes to report the methylation ratio per tested CpG site.

III. TRANSLATING LAB TECHNOLOGIES TO CLINICAL SETTINGS

Efforts applying the commercially available Illumina Infinium Methylation EPIC array (or its previous versions including Infinium HumanMethylation450 Beadchip) to clinical samples have been accomplished.^{70, 71} Such applications indicate the relative maturity of genome-wide DNA methylation detection in a clinical setting. Thus, this section will instead outline research efforts to date for the analysis of protein-DNA interactions with the goal of ultimately translating these technologies into a clinical setting to facilitate precision medicine.

Limitations of ChIP-seq and Variants.

While being the traditional workhorse to probe protein-DNA interactions, limitations of ChIP-seq prevent its routine clinical implementation. One major challenge is its requirement of large sample input. Traditionally, ChIP-seq requires more than 10⁶ cells as starting material per analysis targeting a single mark to ensure effective enrichment of related DNA.^{58, 72} This is unfeasible for clinical biopsies containing fewer than 10³ cells, or other samples of rare cell populations. Furthermore, ChIP-seq is incapable of profiling cellular heterogeneity. Cell lysis and digestion are completed in bulk by combining DNA from all cells, which obscures possible rare signals from small numbers of abnormal or novel subsets within heterogeneous samples. Additionally, ChIP-seq suffers from poor robustness and low reproducibility as it relies largely on operator skills and reagent quality. Moreover, application of this technique to formalin-fixed, paraffin-embedded samples has also been a challenge, and is thus a major focus of current efforts to increase application to archived pathological specimens.^{73, 74} Finally, ChIP-seq is inherently low throughput as it investigates one target per sample (by using a single type of mark-specific antibody per experiment). When combined, these factors complicate the use of conventional ChIP-seq for both routine clinical practice and longitudinal studies on large populations, making it less amenable to profiling a wide range of healthy and abnormal samples.⁷⁵

Promising Approaches of Automated protein-DNA Interaction Analysis.

Recent research efforts have turned to microfluidic tools as powerful alternatives to macroscale platforms for automating various epigenomic assays including ChIP-seq.⁷⁶ Microfluidics features miniaturization of reagent volume, parallelization of multiple samples, and automation of an integrated workflow, which are all especially well-suited to sample-constrained, labor-intensive, and highly operator-dependent epigenetic analyses.⁷⁷ Particularly due to microscale device sizes and features, microfluidics facilitates single cell handling and analysis, making it a fundamentally enabling tool for the investigation of single cell epigenomic heterogeneities confounded by macroscale methods. The toolbox of single cell manipulation and processing on microfluidics has been expanding over the years with the emergence of promising proof-of-concept studies including highly efficient single cell encapsulation in microfluidic droplets (Fig. 4a),⁷⁸ high-throughput single cell reverse transcription PCR (RT-PCR) (Fig. 4b),^{79, 80} and single-cell sequencing with comparable read count and improved sensitivity compared to bulk protocols (Fig. 4c).^{81, 82} Although these methods still face limitations like bias in reaction efficiency associated with minute amount of samples,⁸¹ adapting the capabilities inherent to these powerful tools has enabled many exciting studies aimed at facilitating microfluidic DNA-protein interaction analysis.

Figure 4.

Selected microfluidic tools for epigenomic analysis. (a) Ordered single-bead (top) and single-cell encapsulation (bottom). Scale bars: 100 μm. Adapted with permission from ref. 78. Copyright (2008) The Royal Society of Chemistry. (b) Single cell droplet array RT-PCR. Single cells were encapsulated into droplets with alkaline lysis buffer and incubated to ensure cell lysis. After merging with RT-PCR buffer, droplets containing cells thermocycled, and measured fluorescence intensities were used to analyze targeted RNA. Adapted with permission from ref. 80. Copyright (2018) American Chemical Society. (c) Cells and reagent encapsulation in DNA barcoding-single cell whole transcriptome sequencing with droplets. Scale bar: 100 μm. Adapted with permission from ref. 82. Copyright (2015) Elsevier.

The Quake group pioneered the application of microfluidic techniques to automated, low-input epigenomic testing. They developed automated microfluidic ChIP (AutoChIP) allowing enrichment of ChIP DNA from 2000 cells and demonstrated higher precipitation efficiency than conventional ChIP.^{83, 84} AutoChIP’s multi-layered and valve-actuated ring structures controlled sample loading, bead washing and elution of DNA. By multiplexing AutoChIP structures they obtained a high throughput, automated microfluidic device for ChIP (HTChIP) capable of processing 10,000-cell equivalent chromatin samples in each parallel structure, running 14 ChIP and 2 controls simultaneously (Fig. 5a).⁸⁵ Though AutoChIP and HTChIP improved automation in enrichment and washing, the fixed ring structure volumes limited their applications to a determinate sample size. Moreover, HTChIP did not start directly from whole cells for the assay. This required pre-processing and defined sample volume may limit the applicability for clinical settings, but the work represented an inspiring initial foray into microfluidic ChIP.

Figure 5.

Selected microfluidic devices for epigenomic analysis. (a) The HTChIP platform) for measurements of 16 different targets simultaneously. Adapted with permission from ref. 85. Copyright (2012) The Royal Society of Chemistry. (b) The MOWChIP-seq device for epigenetic profiling with 100 cells. Adapted with permission from ref. 86. Copyright (2015) Springer Nature. (c) An integrated device for on-chip sonication and immunoprecipitation. Adapted with permission from ref. 87. Copyright (2016) American Chemical Society. (d) The Drop-ChIP device integrating cell encapsulation, reagents addition, and DNA barcoding to analyze thousands of cells at single-cell resolution. Adapted with permission from Ref. 91. Copyright (2015) Springer Nature.

The Lu lab has also investigated epigenomic assays for small cell populations by exploiting microfluidic chamber structures. Using a valve-actuated microfluidic chamber to control magnetic bead packing, cell lysate loading and captured DNA elution, they developed microfluidic oscillatory washing–based ChIP-seq (MOWChIP-seq) to capture targeted DNA from 100-cell equivalent chromatin for sequencing and discovered new enhancer regions (Fig. 5b).⁸⁶ Further integration of sonication and immunoprecipitation enabled on-chip shearing of chromatin in addition to DNA capture (Fig. 5c).⁸⁷ They also fabricated multiple microfluidic beds on a single chip to carry out ChIP targeting multiple marks in parallel.^{88, 89} Additionally, by immobilizing antibodies directly to the channel surface for flow-based target capture, they developed SurfaceChIP-seq to investigate histone marks in mouse prefrontal cortex and cerebellum specimens.⁹⁰ These strategies further automated the ChIP workflow; however, MOWChIP-seq’s fixed chamber structure limited the platform’s capability in processing flexible sample sizes (as any changes in sample size would require re-optimization of immunocapture conditions), and SurfaceChIP-seq’s immobilization of antibodies may affect their binding affinity (as the antigen binding sites may be impaired due to potential steric changes). Nonetheless, these devices have improved throughput and multiplexibility while reducing input requirements for microfluidic ChIP.

Besides continuous flow microfluidics, the Weitz and Bernstein labs have applied droplet microfluidics to achieve automated single-cell ChIP-seq (Drop-ChIP) that can handle samples with indeterminate sizes and prepare thousands of cells individually within minutes for subsequent library construction and sequencing.⁹¹ Starting with single cell encapsulation followed by simultaneous introduction of chromatin fragmentation and immunocapture reagents to cell-containing droplets, Drop-ChIP used DNA barcodes to label different cells. Therefore, it was able to analyze a mixture of three different cell types at single-cell resolution and elucidated cell subpopulations (Fig. 5d).⁹¹ The specificity and information content accuracy were high, as ~50% of reads could be aligned to known positive sites, and the coverage of aggregated reads from 50 cells was comparable to conventional profiles. However, the coverage per cell was sparse on the order of 1000 unique reads. Additionally, this device only automated cell lysis, chromatin digestion, and indexing with DNA barcodes; the protocol manually completed immunoprecipitation and ChIP DNA extraction off the device. Furthermore, potential repetitions in DNA barcoding introduced possible false identification of cell types, and thus could confound final analyses.

The Bailey lab recently described a strategy to automate chromatin fragmentation for MNase-seq utilizing droplet microfluidics to process cell samples directly, and their approach yielded high quality nucleosome mapping profiles.⁹² This microfluidic device showed the capability to tolerate different sample sizes, making it amenable toward future clinical practices with variable sample quantities; however, the approach has not yet demonstrated single cell resolution. Promisingly, these and related advances in droplet microfluidic technologies are being translated into mass-manufacturable thermoplastic materials, bringing them closer to at-scale manufacturing and fulfilling their promise for clinical distribution and implementation.^93–98

Despite their respective limitations as early implementations of these technologies, these promising examples show the potential of microfluidics to facilitate semi-automated, low-input (and single-cell) epigenomic studies, and motivate further efforts towards implementation in high-volume clinical settings. When reliable and rapid analysis of patients’ epigenomes becomes clinically routine, the enhanced understanding available will offer insights on the unique combination of epigenomic marks that play crucial roles in disease pathways in individual patients. Thus, future generations of these microfluidic technologies will likely prove to be highly enabling as epigenomics is brought into the realm of precision medicine.

CONCLUSIONS AND OUTLOOK

Over the recent years, increased efforts have aimed at elucidating the epigenomic mechanisms of pathogenesis to provide insights for diagnosis and treatment. While many assays have profiled ncRNA expression, DNA methylation, histone modifications, and other chromatin-regulating molecules, ongoing efforts must focus on translating these basic research techniques to clinical settings. Additionally, single cell epigenome profiling still lags behind single cell genome and transcriptome profiling in terms of commercial platform development and its realization of precision medicine applications. Specifically, ChIP-seq struggles to assess biopsy samples containing small cell numbers due to the aforementioned drawbacks. This limits its application in routine analysis for prognosis, diagnosis and treatment. While microfluidics has provided improved automation and reduced reagent consumption, current microfluidic ChIP approaches still suffer from sample size restrictions due to determinate microfluidic devices, complicated microfluidic operation, and lack of automation of the entire ChIP process from cell input to analytical readout. Additionally, these methods will need to be validated using clinical samples (e.g. tissue biopsies) and compared to bulk-scale approaches to ensure accurate data coverage before their robust adoption into clinical practice. Moreover, characterizing highly dynamic, tissue-specific epigenetic states requires the establishment of reference epigenomes of different tissue cells and across various developmental status.^{75, 99, 100} For diagnostics, it is also critical to enable a rapid and robust version of ChIP-seq and other assays. These difficulties provide a starting point for future studies in microfluidic epigenomic assay development. Encouraged by the potential of initial studies, the application of microfluidics tools to clinical epigenetic profiling promises to fundamentally transform the field of precision medicine.