1 Introduction

Multicellular organisms are composed of highly diverse cell types, all of which originate from fertilized eggs but gradually limit their fates during development, differentiating into cell populations with different functions. Cell fate determination involves a series of complex regulatory events, including changes in gene regulatory networks, reprogramming of epigenetic states, and the role of extracellular signals, etc. Classic developmental biology research has established cell fate maps and lineage relationships. For instance, in model organisms such as nematodes and fruit flies, the origin of each cell line was identified through embryonic tracing. However, the development of higher mammals (such as humans and mice) is more complex, and the key molecular mechanisms in the process of cell fate determination remain to be clarified.

Traditional research methods often rely on the measurement of average signals of large population cells, which may mask significant differences among different cell subpopulations. For instance, in hematopoietic stem cell research, progenitor cells from different subpopulations may be evenly distributed in population sequencing, making it impossible to identify a small number of rapidly proliferating transitional cell populations. In recent years, the rise of single-cell sequencing technology has enabled researchers to analyze the developmental process at the single-cell level. For instance, single-cell RNA sequencing during the early embryonic development stage can reconstruct the differentiation trajectories of different cell types, identify the key branch points determined by early lineages and the accompanying changes in gene expression. These studies have revealed the high asynchrony of the developmental process and the heterogeneity among cells, indicating that cell fate determination does not occur synchronously but that different cells may enter their respective differentiation pathways at different times (Swanson et al., 2021).

Single-cell omics refers to the acquisition and analysis of multi-level information at the single-cell level, including genomic, transcriptomic, epigenomic and even proteomic information. Its rise stems from the urgent need for research on cellular heterogeneity and rare cell types. Traditional high-throughput sequencing requires a large number of cells to be mixed, thereby only providing the average population information and masking the significant differences between different cell states. The development of single-cell sequencing technology enables scientists to "disassemble" tissues and redefine cell classification and state from bottom to top (Wu et al., 2024).

Researchers achieved single-cell transcriptome sequencing for the first time, conducting mRNA sequencing analysis on a single embryonic cell and demonstrating the feasibility of whole-transcriptome analysis at the single-cell level. Since then, with the combination of microfluidic technology and high-throughput sequencing methods, single-cell RNA-seq methods have evolved from low-throughput (such as the SMART seq series based on cell capture) to high-throughput (such as the 10x Genomics platform based on oil droplet microdroplet) (Cao et al., 2022; Pan et al., 2022). At present, each single-cell RNA-seq experiment can simultaneously analyze thousands to tens of thousands of cells, making it possible to construct single-cell maps of complex tissues. For instance, the Cell Atlas Project of Human tissues identified numerous previously undescribed cell types using single-cell RNA sequencing, updating the understanding of tissue cell composition in anatomy textbooks (Hanamsagar et al., 2019; David et al., 2020).

The integrated analysis of single-cell RNA-seq and scATAC-seq data can achieve the comparison of information at the transcriptional level and the epigenetic level, thereby mapping the gene regulatory network at the single-cell scale. This multi-dimensional correlation is the key to understanding the mechanism of cell fate determination. The research of Ma et al. (2020) discovered the so-called "regulated chromatin regions (DORCs)"-a group of open chromatin regions closely related to gene expression-by simultaneously obtaining the gene expression and chromatin accessibility data of each cell. These DORCs often enrich super enhancers, and the changes in their open states precede the expression changes of the corresponding genes, which has been proposed as an indicator of "chromatin potential" for evaluating cell fate transitions. This indicates that the integration of scRNA-seq and scATAC-seq can capture the sequence and causal relationship of epigenetic regulation and transcriptional response during cell differentiation (Lee et al., 2023).

The analysis of integrated data can also identify the core regulatory factors that determine cell fate. On the one hand, scATAC-seq data can provide information on potentially active cis-regulatory elements (such as enhancers). Combined with the scRNA-seq expression data of genes near these elements, it is possible to infer whether a specific enhancer-gene pair may have a regulatory relationship. On the other hand, by calculating the number of transcription factor binding sites (motifs) abundant in the open regions of chromatin in each cell, the possible active transcription factors in each cell can be inferred. Comparing this with the gene expression profile of the same cell can verify whether these inferred transcription factors do indeed play a role at the transcriptional level. For instance, in the integrated analysis of hematopoietic stem cell differentiation, researchers analyzed the scATAC-seq motif enrichment through the chromVAR algorithm and found that the activity changes of key transcription factors specific to lineages (such as GATA1, TAL1, etc.) were consistent with the expression changes of the corresponding genes in scRNA-seq. Such results directly correlate the epigenetic level of factor activity with the transcriptional level of functional output, enhancing the credibility of the role of regulatory factors.

2 Principles and Applications of Single-Cell Omics Technology

2.1 Principles and applications of scRNA-seq technology

The basic principle of single-cell RNA sequencing (scRNA-seq) technology is to reverse transcribe, amplify and sequence intracellular mRNA at the single-cell level, thereby obtaining the gene expression profile of the cell (Khan et al., 2023). To achieve this, it is necessary to address the challenge of having an extremely small initial amount of single-cell RNA (only picked-level RNA), so full-length cDNA amplification or tag sequence amplification methods are usually adopted. Early RNA-SEq methods, such as the Domer strand displacement amplification were capable of capturing full-length transcripts, but their throughput was relatively low. Later developed methods such as Drop-seq and 10x Genomics Chromium encapsulated single cells with microdroplets and added microbeads with cell barcodes, introducing UMI (Unique Molecular Identifier) to quantify mRNA copy numbers. Large-scale parallel sequencing and PCR bias correction have been achieved (Babcock and Weir, 2023). The current mainstream RNA-SEQ protocols typically rely on 3 'end tag sequencing, which can efficiently sequence thousands of cells and obtain their gene expression matrices (Aryamanesh, 2022).

2.2 Application in cell typing and tumor research

A typical application of RNA-SEq is to identify cell types and states in complex tissues. For instance, Travaglini et al. constructed a single-cell transcriptional map of the human lung and identified new subtypes of alveolar cells, etc., expanding the types of human lung cells from over 40 previously to nearly 60 (Rehman et al., 2022). For instance, the application of scRNA-seq in tumor tissues can classify tumor cells into different subpopulations based on gene expression, and simultaneously analyze multiple components such as immune cells and fibroblasts in the tumor microenvironment, providing a basis for studying tumor heterogeneity and immune escape (Saddala et al., 2023). Especially in highly heterogeneous cancers such as breast cancer, scRNA-seq reveals that significant intercellular differences still exist under traditional molecular typing, and drug resistation-related tumor subclones can be discovered (Tian et al., 2018). Therefore, scRNA-seq has been widely applied in numerous fields such as oncology, development, reproduction, and neurology for cell classification, marker identification, pathway enrichment, and pseudo-temporal trajectory analysis, etc. (Khan et al., 2023).

3 Transcriptional and Epigenetic Regulatory Mechanisms of Cell Fate Determination

3.1 Regulation of transcription factors and gene expression

Transcription factors (TF) are proteins that directly bind to DNA to regulate gene transcription and play a core role in cell fate determination (Kumar and Sharma, 2021). The reason why different cell types have distinct functions and morphologies is largely due to the fact that their specific combinations of transcription factors drive different gene expression programs. For example, during the process of muscle cell differentiation, the expression of transcription factors such as myogenic regulatory factors (such as MyoD) can induce the activation of muscle-specific genes, thereby pushing precursor cells towards the fate of muscle cells (Gurdon et al., 2020). Similarly, in neuronal differentiation, the initiation of basic helical-ring-helix transcription factors such as Neurogenin can induce stem cells to transform into neuronal lineages. These examples demonstrate the classic model where a single "master regulator" transcription factor can determine cell fate. However, in a broader context, the determination of cell fate is often the result of the synergistic action of multiple transcription factors (Larcombe et al., 2022). For example, the core transcription factor network that maintains pluripotency of embryonic stem cells includes OCT4, SOX2, NANOG, etc. They mutually regulate and jointly maintain the fate of stem cells. When any one of them is disturbed, differentiation will be triggered (Balsalobre and Drouin, 2022).

3.2 The role of chromatin accessibility in fate determination

Chromatin accessibility refers to the degree of openness of chromatin structure, that is, whether regulatory proteins such as transcription factors can conveniently access DNA sequences (Fan and Huang, 2021). In the process of cell fate determination, the dynamic change of chromatin accessibility is an important prerequisite for gene regulation (Kim et al., 2022). Developmental biology research indicates that there are extensive chromatin remodeling events in the early stage of embryonic development, gradually confining the genome of pluripotent stem cells to specific lineages of epigenetic states.

3.3 Interactive regulation of epigenetics and transcription

A typical example is the formation and function of super-enhancers. These regions are usually occupied by core transcription factors determined by cell identity and recruit large-scale coactivators and chromatin regulatory proteins (Balsalobre and Drouin, 2022). For instance, integrating the data of scRNA-seq and scATAC-seq can reveal that in certain cells, the gene of a key transcription factor simultaneously shows high expression and its promoter/enhancer is highly open, and the binding sites of the factor itself are enriched, suggesting the synergy of its self-regulation and chromatin remodeling (Fan and Huang, 2021).

4 Based on the Key Regulatory Factors Revealed by the Integrated Analysis

4.1 Identification of key transcription factors

Through the integrated analysis of single-cell RNA-seq and ATAC-seq, a direct objective was to identify the key transcription factors that drive cell fate determination (Hamrud et al., 2025). Traditional methods often infer the importance of known factors based on their expression changes. However, integrative analysis can more powerfully determine whether a certain factor is "critical" because it can simultaneously consider the upstream epigenetic state and downstream gene activation of the factor (Zhang et al., 2023).

Specifically, in the integration of data, we can identify candidate key factors from two perspectives: First, if the gene of a certain transcription factor is significantly upregulated during the differentiation process and an open enhancer appears in the chromatin near its locus at the same stage, it indicates that the activation of this factor is permitted by epigenetic regulation and can be suspected as a phylogenetic determinant. For instance, studies on the hematopoietic differentiation of human fetal liver have found that the transcription factor HLF is highly expressed in the early HSC/MPP stage, and its promoter region is specifically open in HSC cells but gradually closes in downstream differentiated cells. This suggests that HLF may play a key role in maintaining the status of stem cells. Further functional experiments also demonstrated that the self-renewal ability of stem cells decreased after HLF knockout. Therefore, the clues initially provided by the integrated analysis were thus verified.

4.2 Cis-regulating element and enhancer interaction

Cis-regulatory elements, including gene promoters, enhancers, silencers, etc., are DNA sequences that determine when and where genes are expressed. Enhancers are particularly important as they can move away from the gene itself and enhance transcription by contacting the promoter through DNA loops. Single-cell ATAC-seq can capture the opening of enhancers on a genome-wide scale, while scRNA-seq tells us the expression of the corresponding genes. Integrating the two can establish enhancer-gene associations and evaluate their significance in cell fate determination (Figure 1) (Finkbeiner et al., 2022).

In integrative analysis, researchers often select open regions close to a certain gene as candidate enhancers based on chromatin accessibility data, and then examine the association between these open regions and gene expression. For instance, Pervolarakis et al. conducted a combined analysis of scRNA and scATAC on mammary epithelial cells and discovered that there were specific open enhancers near some genes related to the maintenance of mammary epithelial cell identity (such as those responsible for lumen/basal surface characteristics). Moreover, the openness of these enhancers is positively correlated with the expression levels of the corresponding genes in different cells. This suggests that these enhancers are likely to directly regulate the corresponding genes, thereby affecting the cell phenotype (Xu et al., 2021).

4.3 Analysis of gene-regulatory networks

By combining the integrated information of key transcription factors and cis-regulatory elements, we can reconstruct the gene regulatory network for cell fate determination. In this kind of network, nodes include transcription factor genes and other regulatory genes, while edges represent regulatory relationships. For instance, transcription factor A regulates the expression of gene B through enhancer X. Integrating single-cell data can help us build more refined and reliable networks.

One example is the previously mentioned study on the development of interneurons in primates. By integrating scRNA and scATAC, the authors constructed a gene regulatory network during development: they paired enhancers that were significantly open at different stages with genes that were significantly expressed, and inferred a number of important factor regulatory relationships. For instance, it was found that the DLX family of transcription factors not only have upregulated expression themselves, but also regulate the enhancer opening of a series of downstream neuronal functional genes near their binding sites, thereby establishing the "DLX regulatory module", which is crucial for the fate of interneurons (Hamrud et al., 2025). At the same time, they also identified some regulatory edges that only exist in primates. For instance, the enhancer of the TH gene has been opened to be associated with the FOXP2 factor, which might imply that the unique FOXP2 regulatory circuit in humans has intervened in the differentiation of certain types of interneurons.

5 Inference of 5 Pathways and Cell Fate Trajectories

One of the key goals of single-cell data analysis is to reconstruct the trajectory of cell differentiation, that is, to infer the sequence of their development or transformation processes based on the continuity of molecular characteristics between cells. Trajectory inference, also known as pseudotime analysis, has many mature methods in scRNA-seq data (such as Monocle, Slingshot, PAGA, etc.) (Durmaz and Scott, 2022). These methods are usually based on measuring the distances between cells in a high-dimensional expression space, and then looking for curves or trees in a low-dimensional space to represent the gradual relationship of cell states. Through trajectory inference, asynchronously differentiated cells can be sorted to identify the starting point, end point and intermediate branch nodes of differentiation, thereby reconstructing the lineage relationship.

Integrating the data of scRNA-seq and scATAC-seq can enhance the accuracy of trajectory inference. One approach is to combine the two types of data to construct a comprehensive low-dimensional embedding space, enabling trajectory inference to take into account both transcriptional and apparent information comprehensively. For example, the weighted Nearest neighbor (WNN) analysis introduced by Seurat v4 can integrate the adjacent information contributed by RNA and ATAC respectively to form a more reliable intercellular relationship map (Figure 2) (Lin et al., 2024). On this basis, trajectory calculation can avoid the pseudo-continuity that may occur solely based on transcription data. Especially in cases where some key transcriptional changes lag behind, ATAC data can provide leading indicators and improve the accuracy of trajectories. For instance, in the differentiation of immune T cells, the transcriptional changes of cell phenotypic transformation may occur after epigenetic changes. It might be difficult to sort with scRNA alone, but after the addition of scATAC, the early epigenetic changes correctly positioned the cells at the front end of the trajectory (Zhang et al., 2024).

6 Actual Case Analysis: Development of the Hematopoietic System or Nervous System

6.1 Integrated analysis of hematopoietic stem cell differentiation

The hematopoietic system provides a classic model for studying the determination of cell fate. Hematopoietic stem cells (HSCS) are located at the top of the differentiation lineage and can generate all types of blood cells, including myeloid (red blood cells, granulocytes, megakaryocytes, etc.) and lymphoid (T cells, B cells, etc.). For a long time, the pathways and regulatory factors of hematopoietic differentiation have attracted much attention. However, there are still unsolved issues regarding the molecular mechanisms of cellular heterogeneity and fate determination at each stage of HSC differentiation. Important progress has been made in the integrated analysis of single-cell RNA-seq and ATAC-seq in this field (Lee et al., 2023).

Ranzoni et al. conducted combined sequencing of scRNA-seq and scATAC-seq on HSC/MPP and subsequent progenitor cells in the liver and bone marrow of human fetuses, analyzing over 8,000 single cells in total. Firstly, based on the scRNA-seq data, they mapped the differentiation trajectories of hematopoietic cells, showing that the HSC/MPP population was at the top of the trajectory and differentiated along three main branches: the red blood cell/megakaryocyte branch, the myeloid mononuclear cell branch, and the lymphocyte branch. Interestingly, they observed significant transcriptional heterogeneity in the HSC/MPP population, suggesting that different transcriptional subpopulations are actually contained within the HSC/MPP defined by surface markers. This provides new evidence for the long-standing debate over whether the HSC group is homogeneous: the HSC is not a single static group but is dynamically composed of multiple subtypes, which may predict different fates.

6.2 Nervous system development and cell type diversity

The development of the nervous system is complex and generates a wide variety of cells, including various types of neurons and glial cells. Take the development of the cerebral cortex as an example. Interneurons (inhibitory neurons) and projection neurons (excitatory neurons) come from different developmental regions, and their fate is determined by delicate spatiotemporal regulation. Single-cell omics integration analysis has also made progress in this field in recent years, such as the study of the development process of interneurons in primate brains (Cai et al., 2025).

After integrating the scATAC data, they explored the changes in chromatin opening and gene regulation at different stages. The results show that many key transcription factors already recognized in rodents (such as mice), such as DLX, NKX2-1, LHX6, etc., function similarly in primates, with their binding sequences significantly open in the chromatin at the corresponding stages. For instance, multiple enhancers of the DLX gene cluster are open at the basal progenitor cell stage, and the DLX gene mRNA is highly expressed, which drives the gradual expression of a series of downstream inhibitory neuron genes (such as GAD1/2, etc.). This indicates that these factors form a conserved regulatory network that drives the fate of interneurons.

6.3 Key findings and biological implications in the case

In hematopoietic cases, integrative analysis revealed the existence of transitional progenitor cell populations (MEMP/GP/LMP) biased towards different lineages in HSC/MPP, enriching the traditional lineage tree model. In neurological cases, multiple intermediate progenitor cell stages have also been identified, such as basal progenitor cells and migrating immature neurons. These transitional states are difficult to capture in population experiments, but single-cell analysis can depict them, indicating that cell fate determination is not achieved in one step but goes through multiple gradual stages.

Both cases show that chromatin opening at the binding sites of key transcription factors often precedes changes in gene expression. For instance, in hematopoiesis, the GATA1 enhancer opens in advance, while in the nervous system, ASCL1 opens its target area first. This validates the "chromatin open leader" hypothesis and emphasizes the role of pioneer transcription factors (Felce et al., 2024). This is of great significance to developmental biology because in the past, most inferences about the chronological order came from indirect reasoning, but now there is direct single-cell evidence to support it.

Integrated analysis not only lists the key factors but also weaves them into network modules. For instance, in hematopoietic cases, there are the HLF-HOXA9 stem cell module and the CEBP-IRF myeloid module; in neurological cases, there are the DLX module and the NEUROD module, etc. These network modules encompass the main regulatory factors and their downstream genes, providing a systematic understanding of the fate determination mechanism rather than acting in isolation. This network perspective is conducive to formulating intervention strategies because regulatory networks often have redundancy and complementarity, and recognition networks are more comprehensive than single genes.

7 Potential Clinical Applications and Disease Mechanism Research

7.1 Tumor heterogeneity and discovery of therapeutic targets

Tumors are typical cases of abnormal determination of cell fate. Normal cells transform into cancer cells with uncontrolled proliferation and abnormal differentiation through a series of gene mutations and epigenetic changes. Within tumors, not only do cancer cells themselves exhibit high heterogeneity, but immune cells and others in the microenvironment are also diverse. This heterogeneity is an important cause of tumor drug resistance and treatment failure. Integrated analysis of single-cell RNA-seq and ATAC-seq provides a powerful tool for analyzing tumor heterogeneity and can identify the characteristics and regulatory networks of different subpopulations within tumors (Raevskiy et al., 2023).

For instance, in the studies of renal cell carcinoma and breast cancer mentioned earlier, integrative analysis revealed that tumor cells can be classified into different "fate" subtypes: one proliferative type, which exhibits high activity in cell cycle genes and the MYC pathway; An invasive type, showing epithelial-mesenchymal transition and AP-1 transcription factor activity. These subtypes may be averaged out at the bulk level, but single-cell technology makes them clearly visible. More importantly, through scATAC-seq, the key cis-elements and transcription factors driving each subtype can be identified. This suggests that different subgroups of tumor cells may require different treatment strategies.

7.2 Immune cell fate and disease mechanisms

The immune system is another system with highly malleable cellular fate. Single-cell omics integration technology provides a new approach for studying the differentiation fate and disease mechanisms of immune cells. In cases of chronic infection or cancer, T cells often head towards functional failure. The trajectory and regulatory network of this exhausted state were revealed through the integration of scRNA+scATAC analysis. Research has found that the transcription factor TOX is gradually upregulated on the depletion trajectory and opens the chromatin regions of inhibitory genes such as PDCD1 (Leslie, 2020). In autoimmune diseases such as systemic lupus erythematosus (SLE), studies have identified an inflammation-related B-cell intermediate state, characterized by the expression of T-bet and being in a transitional state between memory and plasma cells. Such abnormal fate decisions may be the key link in the occurrence of diseases.

7.3 Prospects of regenerative medicine and precision medicine

The core of regenerative medicine lies in controlling the fate of stem cells or somatic cells to transform them into the desired functional cells. Integrated single-cell RNA/ATAC analysis provides key support for this field (Wang et al., 2024). In in vitro induction differentiation, single-cell analysis can identify intermediate states and deviated pathways. For instance, it can diagnose the causes of differentiation failure by discovering that key enhancers are not open and repair them through small molecule regulation. Meanwhile, non-target lineage cells mixed in can be monitored to provide a basis for purification. In the field of precision medicine, abnormal fateful determination nodes can be identified through single-cell integrated sequencing of biopsy tissues. For instance, in the brain tissues of patients with neurodegenerative diseases, observing the arrest of oligodendrocyte precursors can indicate myelin formation disorders, thereby providing directions for drug intervention.

8 Current Limitations and Future Development Directions

scRNA-seq usually fails to detect low-abundance transcripts, which results in the omission of some key regulatory factors, such as low-expressed transcription factors. Meanwhile, scATAC-seq has only several hundred thousand DNA sequencing reads per cell covering the entire genome, resulting in a large number of "zero" signals and random noise. These technical noises increase the difficulty of integration analysis and are prone to false positive or false negative associations. Although methods such as high-load sequencing and cell merging analysis have been developed in recent years to improve the signal-to-noise ratio, the fundamental solution requires innovations in sequencing chemistry and instrumentation. For instance, developing library preparation methods with lower amplification bias or directly reading epigenetic markers using nanopore sequencing may enhance sensitivity and accuracy.

Secondly, single-cell multi-omics sequencing is still relatively complex and expensive at present, and the data output rate is not high. In integrative analysis, we usually measure scRNA and scATAC separately and then pair the cells using computational methods. This kind of matching may have incorrect links because different cell patterns do not necessarily correspond one-to-one. Ideally, RNA and ATAC information can be obtained simultaneously at the single-cell level. Currently, although the Multiome kit provided by 10x Genomics can measure two types of data of the same cell, its cost is several times that of the single-mode, and the data quality is also slightly compromised. In the future, more efficient simultaneous sequencing platforms will be needed, which can even measure DNA methylation, protein expression, etc. at the same time, to achieve true panoramic single-cell analysis.

With the advancement of technology, the types of omics we can measure are constantly increasing. Besides transcription and accessibility, there are also DNA methylation, histone modification, proteins, metabolites, and so on. How to integrate such multi-level information is a huge challenge in the future. Most of the existing analytical methods deal with bimodal, and the complexity increases exponentially if there are more. Moreover, the time scales of signals in different modes vary. For instance, transcription can be at the minute level, while methylation changes may be at the hour or day level. How to consider time lag in analysis? In addition, the quality of multimodal data is uneven. Some omics techniques are very mature (RNA), while others are very new and noisy (proteomics such as CITE-seq). Simply piecing them together may be dominated by noise. It is necessary to develop a new computing framework, perhaps introducing physical models or graph models to fit the relationships at each layer.

Another frontier is the spatiotemporal dimension. Cell fate determination often occurs in specific tissue structures and at specific developmental times. The classic method of single-cell sequencing disintegrates tissues and loses spatial information. Sequencing also damages cells and makes it impossible to track time. Although spatial transcriptomics and lineage tracing and other methods have emerged, integrating these with conventional single-cell data remains a new challenge. For instance, how can spatial position constraints be introduced into cell clustering or trajectory inference? How to correlate the lineage tree with the transcriptional trajectory? Currently, attempts have been made to integrate spatial and single-cell data through algorithms such as SpaceFlow, but they are still in the initial stage. In the future, it is hoped to establish a 4D (spatial 3D+ temporal) cell fate model, which requires a dual approach of experiments and computations: experimentally, information is obtained through time point sampling, lineage tracing, and in vivo imaging; computationally, dynamic network modeling, spatio-temporal point process analysis, etc. are developed for integration. If successful, we will be able to "see" how cells in a living embryo determine their fate and when and where they send out what signals-this will be a dream come true for developmental biology.

As sequencing costs further decrease and operations become simplified, single-cell sequencing will become a standard feature in many biomedical studies. It might be as widespread as PCR is now. At that time, a vast amount of data will emerge, and the community needs to establish a unified data storage and analysis platform to achieve data sharing and reuse. Standardized analysis procedures and quality control will also be established to make the results from different laboratories comparable.

Research on single-cell fate determination will be deeply integrated with fields such as physics, engineering, and clinical practice. Physically, analogical phase transition theory and stochastic process theory can help understand the randomness and certainty of cell state transitions. In engineering, microfluidics, high-dimensional data visualization, and AI algorithms will continuously inject new vitality. Clinically, the accumulation of a large number of single-cell data from patient samples will give rise to a new concept of "single-cell pathology", and doctors may be able to diagnose and classify diseases based on single-cell maps. For example, distinguishing different tumor subtypes and determining treatment plans, etc. Single-cell technology may be applied in sensitive fields such as reproduction and development, for instance, in the study of the fate of cells before human embryo implantation, which involves ethics. Society should establish norms for such research. At the same time, when technology is applied to human enhancement (such as optimizing stem cells for anti-aging), the impact also needs to be evaluated.

Acknowledgments

We would like to thank Mr. Jiao continuous support throughout the development of this study.

Conflict of Interest Disclosure

The authors affirm that this research was conducted without any commercial or financial relationships that could be construed as a potential conflict of interest.

References

Aryamanesh N., 2022, A reproducible and dynamic workflow for analysis and annotation of scRNA-seq data, In: Epiblast stem cells: methods and protocols, New York, NY: Springer US, pp.101-140.

https://doi.org/10.1007/978-1-0716-2281-0_10

Babcock B.R., and Weir S., 2023, scRNA-seq for microcephaly research [I]: single-cell droplet encapsulation, mRNA capture, and cDNA synthesis, In: Microcephaly: methods and protocols, New York, NY: Springer US, pp.83-97.

https://doi.org/10.1007/978-1-0716-2752-5_8

Balsalobre A., and Drouin J., 2022, Pioneer factors as master regulators of the epigenome and cell fate, Nature Reviews Molecular Cell Biology, 23(7): 449-464.

https://doi.org/10.1038/s41580-022-00464-z

Cai L., Ma X., and Ma J., 2025, Integrating scRNA-seq and scATAC-seq with inter-type attention heterogeneous graph neural networks, Briefings in Bioinformatics, 26(1): bbae711.

https://doi.org/10.1093/bib/bbae711

Cao J., Chen X., Huang S., Shi W., Fan Q., Gong Y., Peng Y., Wu L., and Yang C., 2022, Microfluidics-based single cell analysis: from transcriptomics to spatiotemporal multi-omics, TrAC Trends in Analytical Chemistry, 158: 116868.

https://doi.org/10.1016/j.trac.2022.116868

David F.P.A., Litovchenko M., Deplancke B., and Gardeux V., 2020, ASAP 2020 update: an open, scalable and interactive web-based portal for (single-cell) omics analyses, Nucleic Acids Research, 48(W1): W403-W414.

https://doi.org/10.1093/nar/gkaa412

Durmaz A., and Scott J., 2022, Stability of scRNA-seq analysis workflows is susceptible to preprocessing and is mitigated by regularized or supervised approaches, Evolutionary Bioinformatics, 18: 11769343221123050.

https://doi.org/10.1177/11769343221123050

Fan T., and Huang Y., 2021, Accessible chromatin reveals regulatory mechanisms underlying cell fate decisions during early embryogenesis, Scientific Reports, 11(1): 7896.

https://doi.org/10.1038/s41598-021-86919-3

Felce C., Gorin G., and Pachter L., 2024, A biophysical model for ATAC-seq data analysis, bioRxiv, 2024: 577262.

https://doi.org/10.1101/2024.01.25.577262

Finkbeiner C.R., Ortuño-Lizarán I., Sridhar A., Hooper M., Petter S., and Reh T.A., 2022, Single-cell ATAC-seq of fetal human retina and stem-cell-derived retinal organoids shows changing chromatin landscapes during cell fate acquisition, Cell Reports, 38(4): 110294.

https://doi.org/10.1016/j.celrep.2021.110294

Hamrud E., Leese J., Thiery A.P., Buzzi A.L., Vigilante A., Briscoe J., and Streit A., 2025, A single cell ATAC-seq atlas uncovers dynamic changes in chromatin accessibility during cell fate specification at the neural plate border, bioRxiv, 2025: 652017.

https://doi.org/10.1101/2025.05.03.652017

Hanamsagar R., Marcus R., Chamberlain M., de Rinaldis E., and Savova V., 2019, Optimum processing conditions for single cell RNA sequencing on frozen human PBMCs, The Journal of Immunology, 202(1_Supplement): 131.15.

https://doi.org/10.4049/jimmunol.202.supp.131.15

Khan S.U., Huang Y., Ali H., Ali I., Ahmad S., Khan S., Hussain T., Ullah M., and Lu K., 2023, Single-cell RNA sequencing (scRNA-seq): advances and challenges for cardiovascular diseases, Current Problems in Cardiology, 49(2): 102202.

https://doi.org/10.1016/j.cpcardiol.2023.102202

Kim E.D., Dorrity M.W., Fitzgerald B.A., Seo H., Sepuru K.M., Queitsch C., Mitsuda N., Han S., and Torii K.U., 2022, Dynamic chromatin accessibility deploys heterotypic cis/trans-acting factors driving stomatal cell-fate commitment, Nature Plants, 8(12): 1453-1466.

https://doi.org/10.1038/s41477-022-01304-w

Kumar R., and Sharma A., 2021, Transcription factor stoichiometry in cell fate determination, Journal of Genetics, 100(2): 27.

https://doi.org/10.1007/s12041-021-01278-2

Larcombe M., Hsu S., Polo J., and Knaupp A.S., 2022, Indirect mechanisms of transcription factor-mediated gene regulation during cell fate changes, Advanced Genetics, 3(4): 2200015.

https://doi.org/10.1002/ggn2.202200015

Lee M., Kaestner K., and Li M., 2023, Benchmarking algorithms for joint integration of unpaired and paired single-cell RNA-seq and ATAC-seq data, Genome Biology, 24(1): 244.

https://doi.org/10.1101/2023.02.01.526609

Leslie C., 2020, Chromatin state and scRNA-seq analysis defines a common differentiation trajectory towards T-cell exhaustion, In: Proceedings of the AACR special conference on tumor immunology and immunotherapy, Cancer Immunology Research, 8(3 Suppl): IA20.

https://doi.org/10.1158/2326-6074.TUMIMM19-IA20

Lin Y., Wu T.Y., Chen X., Wan S., Chao B., Xin J., Yang J., Wong W.H., and Wang Y., 2024, scTIE: data integration and inference of gene regulation using single-cell temporal multimodal data, Genome Research, 34: 119-133.

https://doi.org/10.1101/gr.277960.123

Ma S., Zhang B., LaFave L.M., Earl A.S., Chiang Z., Hu Y., Ding J., Brack A., Kartha V., Tay T., Law T., Lareau C., Hsu Y., Regev A., and Buenrostro J.D., 2020, Chromatin potential identified by shared single-cell profiling of RNA and chromatin, Cell, 183(4): 1103-1116. e20.

https://doi.org/10.1016/j.cell.2020.09.056

Pan Y., Cao W., Mu Y., and Zhu Q., 2022, Microfluidics facilitates the development of single-cell RNA sequencing, Biosensors, 12(7): 450.

https://doi.org/10.3390/bios12070450

Raevskiy M., Yanvarev V., Jung S., del Sol A., and Medvedeva Y., 2023, Epi-Impute: single-cell RNA-seq imputation via integration with single-cell ATAC-seq, International Journal of Molecular Sciences, 24(7): 6229.

https://doi.org/10.3390/ijms24076229

Rehman A., Rashid A., and Anwar I., 2022, Single cell RNA sequencing (scRNA-seq) as an emerging technology in cancer research, Proceedings of the Pakistan Academy of Sciences: B. Life and Environmental Sciences, 58(3): 19-28.

https://doi.org/10.53560/ppasb(58-3)663

Saddala M., Mundla S., Patyal N., and Dash S., 2023, Single-cell RNA sequencing (scRNA-seq) data analysis of retinal homeostasis and degeneration of microglia, In: Diabetic retinopathy: methods and protocols, New York, NY: Springer US, pp.91-106.

https://doi.org/10.1007/978-1-0716-3255-0_6

Swanson E., Lord C., Reading J., Heubeck A., Genge P., Thomson Z.,Weiss M., Li X., Savage A., Green R., Torgerson T., Bumol T., Graybuck L., and Skene P., 2021, Simultaneous trimodal single-cell measurement of transcripts, epitopes, and chromatin accessibility using TEA-seq, eLife, 10: e63632.

https://doi.org/10.7554/eLife.63632

Tian L., Dong X., Freytag S., Lê Cao K., Su S., Amann-Zalcenstein D., Weber T., Seidi A., Naik S., and Ritchie M., 2018, scRNA-seq mixology: towards better benchmarking of single cell RNA-seq protocols and analysis methods, bioRxiv, 2018: 433102.

https://doi.org/10.1101/433102

Wang H., Liang M., Wu J., Peng X., Zhu K., and Zheng Z., 2024, Strategies of integrated analysis of ATAC-seq and RNA-seq data, Science of Advanced Materials, 16(1): 130-140.

https://doi.org/10.1166/sam.2024.4599

Wu Y., Yang X., Dai Y., Zhao Z., Zhu J., Guo H., and Yang R., 2024, Single-cell sequencing to multi-omics: technologies and applications, Biomarker Research, 12(1): 110.

https://doi.org/10.1186/s40364-024-00643-4

Xu K., Zhang W., Wang C., Hu L., Wang R., Wang C., Tang L., Zhou G., Zou B., Xie H., Tang J., and Guan X., 2021, Integrative analyses of scRNA-seq and scATAC-seq reveal CXCL14 as a key regulator of lymph node metastasis in breast cancer, Human Molecular Genetics, 30(5): 370-380.

https://doi.org/10.1093/hmg/ddab042

Zhang R., Qiu C., Filippova G., Li G., Shendure J., Vert J., Deng X., Disteche C., and Noble W., 2024, Multi-condition and multi-modal temporal profile inference during mouse embryonic development, bioRxiv, 2024: 583179.

https://doi.org/10.1101/2024.03.03.583179

Zhang S., Pyne S., Pietrzak S.J., Halberg S., McCalla S.G., Siahpirani A.F., Sridharan R., and Roy S., 2023, Inference of cell type-specific gene regulatory networks on cell lineages from single cell omic datasets, Nature Communications, 14(1): 3064.

https://doi.org/10.1038/s41467-023-38637-9