Jump to content

Single-cell multi-omics integration

Add links
From Wikipedia, the free encyclopedia
This is an old revision of this page, as edited by Shortytot (talk | contribs) at 21:34, 22 February 2024 (→Mixed Integration). The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.

Single-cell multi-omics integration describes a suite of computational methods used to harmonize information from multiple "omes" to jointly analyze biological phenomena[1][2][3][4]. This approach allows researchers to discover intricate relationships between different chemical-physical modalities by drawing associations across various molecular layers simultaneously. Multi-omics integration approaches can be categorized into four broad categories: Early integration, intermediate integration, late integration, and mixed integration methods[5]. Multi-omics integration can enhance experimental robustness by providing independent sources of evidence to address hypotheses, leveraging modality-specific strengths to compensate for another's weaknesses through imputation, and offering cell-type clustering and visualizations that are more aligned with reality[1][2].

Background

The emergence of single-cell sequencing technologies has revolutionized our understanding of cellular heterogeneity, uncovering a nuanced landscape of cell types and their associations with biological processes. Single-cell omics technologies has extended beyond the transcriptome to profile diverse physical-chemical properties at single-cell resolution, including whole genomes/exomes, DNA methylation, chromatin accessibility, histone modifications, epitranscriptome (e.g., mRNAs, microRNAs, tRNAs, lncRNAs), proteome, phosphoproteome, metabolome, and more[3][6][7]. In fact, there is an expanding repository of publicly available single-cell datasets, exemplified by growing databases such as the Human Cell Atlas Project (HCA), the Cancer Genome Atlas (TCGA), and the ENCODE project[8][9][10][11][12]. With the increasing diversity in both available datasets and data types, multi-omics data integration and multimodal data analysis represent pivotal trajectories for the future of systems biology.

Single-cell multi-omics integration can reveal underappreciated relationships between chemical-physical modalities, broaden our definition of cell states beyond single modality feature profiles, and provide independent evidence during analysis to support testing of biological hypotheses. However, the high dimensionality (features > observations), high degree of stochastic technical and biological variability, and sparsity of single-cell data (low molecule recovery efficiency) make computational integration a challenging problem[13][14][15][16]. Furthermore, different solutions for multi-omics integration are available depending on factors such as whether the data is matched (simultaneous measurements derived from the same cell) or unmatched (measurements derived from different cells), whether cell-type annotations are available, or whether modality feature conversion is available, with different implementations tailored to suit the specific use case[1]. As such, there are multiple approaches to single-cell data integration, each with a distinct use case, and each with its own set of advantages and disadvantages[1][5][17].

Methodology

Early Integration

Early integration involves concatenating two or more omic datasets (eg. scRNA-seq data and scATAC-seq data) into a single merged data matrix. Despite the advantages of simplicity and being able to consider dependencies between features, the inherent nature of concatenating two datasets together results in differing dimensions and scales among features. More importantly, the resulting matrix would become an even higher dimensional dataset (hence dimensionality reduction is often necessary). To mitigate these issues, strategies like feature selection and dimensionality reduction (eg. PCA, CCA, NMF) are employed - and as mentioned earlier, is often necessary. Regardless, due to these challenges, early data integration has most commonly been used to concatenate different datasets of the same datatype (eg. Integrating two different scRNA-seq datasets).

Intermediate Integration

Intermediate integration strategies aim to analyze multiple omic datasets at the same time without the need for data transformation prior to analysis. The main approaches to doing so include similarity-based integration, joint dimension reduction, and statistical modeling.

Similarity-based integration involves identifying similarities or patterns across multi-omic datasets through the use of spectral clustering (eg. Spectrum and PC-MSC) which cluster cells based on similarity matrices derived from multi-omic datasets or graph fusion algorithms (eg. Seurat4) which construct graphs from individual omics layers and merge them into a single graph.

In joint dimension reduction, the aim is to reduce the complexity of the multi-omics data by projecting them into a lower dimensional latent space such that the different omics layers can be compared and analyzed together. Canonical correlation analysis (CCA), non-negative matrix factorization (NMF) and manifold alignment are common methods for doing joint dimensionality reduction. Tools that use CCA and its extension, sparse CCA, such as Seurat3 and bindSC identifies linear relationships between datasets by finding linear combinations of variables that maximize their correlations with one another. Tools that use NMF (eg. LIGER and coupledNMF) extracts low-dimensional representations of high-dimensional data such that shared and dataset-specific factors across the multiple omics datasets can be identified. Manifold alignment (eg., MATCHER and MAGAN) refers to an approach where a lower dimension representation of the multi-omic datasets are created individually and then aligned in a common latent space.

Statistical approaches can also be used to integrate information from multi-omic datasets. One well known approach is the Bayesian framework which facilitates probabilistic modeling of the multi-omic datasets. Tools that use a Bayesian clustering framework such as BREM-SC can jointly cluster multi-omic datasets and identify cell clusters. Another tool that uses a Bayesian approach to conduct multi-omic integration is Clonealign which as the name suggests, is able to integrate gene expression and copy number profiles to study cancer clones.

Late Integration

Late integration refers to the straightforward approach of processing and modeling each omics dataset separately, then combining the two models at the very end. The advantage of this, lies in the fact that there are well-established tools already designed for each omics modality as different clustering algorithms may be tailored to different omics data types. While late integration approaches have been commonly used in the context of bulk multi-omics studies (eg., Cluster-of-clusters analysis and Kernel Learning Integrative Clustering), late integration in the context of single cell experiments is still a rapidly evolving field. One method of single cell multi-omics late integration known as ensemble clustering (eg. SAME-clustering, Sc-GPE, EC-PGMGR), have demonstrated promising potential in aggregating clustering results from diverse sources. It combines the clustering results from different omics datasets and creates robust consensus clustering which models the relationships between the individual clustering results to find an improved global clustering solution across the different modalities.

However, while late integration is a good solution to handling single-cell multi-omics datasets, it inherently lacks the capability to capture interactions and relationships between different omics modalities. The whole point of multi-omics integration lies in its ability to effectively analyze the inter-omics relationships present in multi-omics data, enabling us to better understand the underlying biological mechanisms driving disease pathogenesis. Hence, while late integration strategies have their merits, it essentially is just single-omics analysis done on multiple datatypes which is not necessarily multi-omics integration.

Dimensionality Reduction

(Link to Dimensionality Reduction Wikipedia page)

Dimensionality reduction refers to the transformation of high dimensional data into a lower dimensional dataset. This decrease in dimensionality reduces noise and simplifies the dataset, resulting in easier handling of data. Dimensionality reduction can be conducted using either feature selection or feature extraction. The former takes the original omic layers and retains only the variables that are important while the latter transforms the original input features into combinations of the original features. Dimensionality reduction is often a necessity especially in the context of a high dimensional dataset and if a particular integration strategy requires it (eg. early and intermediate integration).

Considerations of Data Integration

Noise

dfasdf

Dataset Compatibility

safasdfdsfds

Dimensionality

dfsdfadfds

Oversimplification of Modality Mapping

dsafsfdfa

Interpretability and Validation

asfdfsfd

Matched and Unmatched Data

dfdasfasfdasf

Applications and Uses

While single-modality datasets have proven to be a mainstay in systems biology, combining biological information across multiple modalities has the potential to address biological questions that cannot be inferred by a single data type alone. For example, the integration of transcriptome and DNA accessibility has enabled the development of bioinformatic tools to infer cell-type-specific gene regulatory networks[18][19][20]. This is achieved by leveraging transcription factor and target gene expression along with cis-regulatory information to impute relevant transcription factors and their regulatory partners. Another application for multi omics integration is in expanding definitions of cell states incorporating features observed across multiple modalities. For instance, integrating protein marker detection with transcriptome profiling using a multi-omics sequencing technology such as CITE-seq can resolve cell state signatures based on joint gene regulatory and surface marker expression[21]. This enables more robust inferences regarding cellular phenotypes, which are akin to and directly comparable with results from classical flow cytometry. Moreover, defining cell states based on clustering analysis within an integrated latent space may offer more stable estimations of cellular phenotypes compared to analysis within a single-modality latent space[1]. Furthermore, multi omics integration can overcome modality-specific limitations. For example, most spatial transcriptomic sequencing technologies suffer from limited spatial resolution (pixels comprising a mixture of local cells) and low feature complexity[22]. Integration of spatial transcriptomics with scRNAseq can help overcome these limitations by supporting the spatial deconvolution of low-resolution readouts and estimating the frequencies of each cell type[23][24].

References

  1. ^ a b c d e Miao, Zhen; Humphreys, Benjamin D; McMahon, Andrew P; Kim, Junhyong (2021). "Multi-omics integration in the age of million single-cell data". Nat Rev Nephrol. 17 (11): 710–724. doi:10.1038/s41581-021-00463-x. PMC 9191639. PMID 34417589.
  2. ^ a b Subramanian, Indhupriya (2020). "Multi-omics Data Integration, Interpretation, and Its Application". Bioinform Biol Insights. 14. doi:10.1177/1177932219899051. PMID 32076369.
  3. ^ a b Stuart, Tim; Sajita, Rahul (2019). "Integrative single-cell analysis". Nat Rev Genet. 20 (5): 257–272. doi:10.1038/s41576-019-0093-7. PMID 30696980. S2CID 59409752.
  4. ^ Li, Yunjin; Ma, Lu; Wu, Duojiao; Chen, Geng (2021). "Advances in bulk and single-cell multi-omics approaches for systems biology and precision medicine". Brief Bioinform. 22 (5). doi:10.1093/bib/bbab024. PMID 33778867.
  5. ^ a b Adossa, Nigatu; Khan, Sofia; Rytkönen, Kalle T; Elo, Laura L (2021). "Computational strategies for single-cell multi-omics integration". Comput Struct Biotechnol J. 19: 2588-2596. doi:10.1016/j.csbj.2021.04.060.
  6. ^ Baysoy, Alev; Bai, Zhiliang; Satija, Rahul; Fan, Rong (2024). "The technological landscape and applications of single-cell multi-omics". Nat Rev Mol Cell Biol. 24 (10): 695–713. doi:10.1038/s41580-023-00615-w. PMC 10242609. PMID 37280296.
  7. ^ Macaulay, Iain C; Ponting, Chris P; Voet, Thierry (2017). "Single-Cell Multiomics: Multiple Measurements from Single Cells". Trends Genet. 33 (2): 155-168. doi:10.1016/j.tig.2016.12.003.
  8. ^ Regev, Aviv; Teichmann, Sarah A; Lander, Eric S; Amit, Ido; Benoist, Christophe; Birney, Ewan; Bodenmiller, Bernd; Campbell, Peter; Carninci, Piero; Clatworthy, Menna; Clevers, Hans; Deplancke, Bart; Dunham, Ian; Eberwine, James; Eils, Roland; Enard, Wolfgang; Farmer, Andrew; Fugger, Lars; Göttgens, Berthold; Hacohen, Nir; Haniffa, Muzlifah; Hemberg, Martin; Kim, Seung; Klenerman, Paul; Kriegstein, Arnold; Lein, Ed; Linnarsson, Sten; Lundberg, Emma; Lundeberg, Joakim; Majumder, Partha; Marioni, John C; Merad, Miriam; Mhlanga, Musa; Nawijn, Martijn; Netea, Mihai; Nolan, Garry; Pe'er, Dana; Phillipakis, Anthony; Ponting, Chris P; Quake, Stephen; Reik, Wolf; Rozenblatt-Rosen, Orit; Sanes, Joshua; Satjia, Rahul; Schumacher, Ton N; Shalek, Alex; Shapiro, Ehud; Sharma, Padmanee; Shin, Jay W; Stegle, Oliver; Stratton, Michael; Stubbington, Michael J T; Theis, Fabian J; Uhlen, Matthias; Van Oudenaarden, Alexander; Wagner, Allon; Watt, Fiona; Weissman, Jonathan; Wold, Barbara; Xavier, Ramnik; Yosef, Nir (2017). "The Human Cell Atlas". eLife. 6. doi:10.7554/eLife.27041. PMC 5762154. PMID 29206104.
  9. ^ Lindeboom, Rik G.H; Regev, Aviv; Teichmann, Sarah A (2021). "Towards a Human Cell Atlas: Taking Notes from the Past". Trends Genet. 37 (7): 625–630. doi:10.1016/j.tig.2021.03.007. PMID 33879355.
  10. ^ Weinstein, John N; Collisson, Eric A; Mills, Gordon B; Shaw, Kenna R Mills; Ozenberger, Brad A; Ellrott, Kyle; Shmulevich, Ilya; Sander, Chris; Stuart, Joshua M (2013). "The Cancer Genome Atlas Pan-Cancer analysis project". Nat Genet. 45 (10): 1113-1120. doi:10.1038/ng.2764. PMID 24071849.
  11. ^ The ENCODE Project Consortium (2012). "An integrated encyclopedia of DNA elements in the human genome". Nature. 489 (7414): 57–74. doi:10.1038/nature11247. PMC 3439153. PMID 22955616.
  12. ^ The ENCODE Project Consortium (2020). "Expanded encyclopaedias of DNA elements in the human and mouse genomes". Nature. 583 (7818): 699-710. doi:10.1038/s41586-020-2493-4. PMID 32728249.
  13. ^ Lähnemann, David; Köster, Johannes; Szczurek, Ewa; McCarthy, Davis J; Hicks, Stephanie C; Robinson, Mark D; Vallejos, Catalina A; Campbell, Kieran R; Beerenwinkel, Niko; Mahfouz, Ahmed; Pinello, Luca; Skums, Pavel; Stamatakis, Alexandros; Attolini, Camille Stephan-Otto; Aparicio, Samuel; Baaijens, Jasmijn; Balvert, Marleen; Barbanson, Buys De; Cappuccio, Antonio; Corleone, Giacomo; Dutilh, Bas E; Florescu, Maria; Guryev, Victor; Holmer, Rens; Jahn, Katharina; Lobo, Thamar Jessurun; Keizer, Emma M; Khatri, Indu; Kielbasa, Szymon M; Korbel, Jan O; Kozlov, Alexey M; Kuo, Tzu-Hao; Lelieveldt, Boudewijn P.F; Mandoiu, Ion I; Marioni, John C; Marschall, Tobias; Mölder, Felix; Niknejad, Amir; Rączkowska, Alicja; Reinders, Marcel; Ridder, Jeroen De; Saliba, Antoine-Emmanuel; Somarakis, Antonios; Stegle, Oliver; Theis, Fabian J; Yang, Huan; Zelikovsky, Alex; McHardy, Alice C; Raphael, Benjamin J; Shah, Sohrab P; Schönhuth, Alexander (2020). "Eleven grand challenges in single-cell data science". Genome Biol. 21 (1): 31. doi:10.1186/s13059-020-1926-6. PMC 7007675. PMID 32033589.
  14. ^ Santiago-Rodriguez, Tasha M; Hollister, Emily B (2021). "Multi 'omic data integration: A review of concepts, considerations, and approaches". Semin Perinatol. 45 (6). doi:10.1016/j.semperi.2021.151456. PMID 34256961. S2CID 235822759.
  15. ^ Yuan, Guo-Cheng; Cai, Long; Elowitz, Michael; Enver, Tariq; Fan, Guoping; Guo, Guoji; Irizarry, Rafael; Kharchenko, Peter; Kim, Junhyong; Orkin, Stuart; Quackenbush, John; Saadatpour, Assieh; Schroeder, Timm; Shivdasani, Ramesh; Tirosh, Itay (2017). "Challenges and emerging directions in single-cell analysis". Genome Biol. 18 (1): 84. doi:10.1186/s13059-017-1218-y. PMC 5421338. PMID 28482897.
  16. ^ Argelaguet, RICARD; Cuomo, Anna S. E; Stegle, Oliver; Marioni, John C (2021). "Computational principles and challenges in single-cell data integration". Nat Biotechnol. 39 (10): 1202-1215. doi:10.1038/s41587-021-00895-7. PMID 33941931. S2CID 233722751.
  17. ^ Wu, Yan; Zhang, Kun (2020). "Tools for the analysis of high-dimensional single-cell RNA sequencing data". Nat Rev Nephrol. 16 (7): 408-421. doi:10.1038/s41581-020-0262-0. S2CID 214672522.
  18. ^ Kim, Daniel; Tran, Andy; Kim, Hani Jieun; Lin, Yingxin; Yang, Jean Yee Hwa; Yang, Pengyi (2023). "Gene regulatory network reconstruction: harnessing the power of single-cell multi-omic data". npj Syst Biol Appl. 9 (1): 51. doi:10.1038/s41540-023-00312-6. PMC 10587078. PMID 37857632.
  19. ^ Bravo González-Blas, Carmen; De Winter, Seppe; Hulselmans, Gert; Hecker, Nikolai; Matetovici, Irina; Christiaens, Valerie; Poovathingal, Suresh; Wouters, Jasper; Aibar, Sara; Aerts, Stein (2023). "SCENIC+: single-cell multiomic inference of enhancers and gene regulatory networks". Nat Methods. 20 (9): 1355–1367. doi:10.1038/s41592-023-01938-4. PMC 10482700. PMID 37443338.
  20. ^ Fleck, Jonas Simon; Jansen, Sophie Martina Johanna; Whollny, Damian; Zenk, Fides; Seimiya, Makiko; Jain, Akanksha; Okamoto, Ryoko; Santel, Malgorzata; He, Zhisong; Camp, J. Gray; Treutlein, Barbara (2023). "Inferring and perturbing cell fate regulomes in human brain organoids". Nature. 621 (7978): 675-372. doi:10.1038/s41586-022-05279-8. PMID 36198796.
  21. ^ Stoeckius, Marlon; Hafemeister, Christoph; Stephenson, William; Houck-Loomis, Brian; Chattopadhyay, Pratip K; Swerdlow, Harold; Sajita, Rahul; Smibert, Peter (2017). "Simultaneous epitope and transcriptome measurement in single cells". Nat Methods. 14 (9): 865–868. doi:10.1038/nmeth.4380. PMC 5669064. PMID 28759029.
  22. ^ Atta, Lyla; Fan, Jean (2021). "Computational challenges and opportunities in spatially resolved transcriptomic data analysis". Nat Commun. 12 (1): 5283. doi:10.1038/s41467-021-25557-9. PMC 8421472. PMID 34489425.
  23. ^ Andersson, Alma; Bergenstråhle, Joseph; Asp, Michaela; Jurek, Aleksandra; Fernández Navarro, José; Lundeberg, Joakim (2020). "Single-cell and spatial transcriptomics enables probabilistic inference of cell type topography". Commun Biol. 3 (1): 565. doi:10.1038/s42003-020-01247-y. PMID 33037292.
  24. ^ Ma, Ying; Zhou, Xiang (2022). "Spatially informed cell-type deconvolution for spatial transcriptomics". Nat Biotechnol. 40 (9): 1349-1359. doi:10.1038/s41587-022-01273-7. PMID 35501392.
Retrieved from "https://en.wikipedia.org/w/index.php?title=Single-cell_multi-omics_integration&oldid=1209632481"