The PIIO faculty and staff will concentrate on the study of molecular and cellular mechanisms of immune tolerance and cancer immune evasion in preclinical human-relevant cancer models and patients with cancer. At the same time, they will work to develop first-in-class, first-in-man clinical IO agents and advance clinical IO with groundbreaking ideas. The ultimate goal is to invent IO agents/platforms that improve care for all patients with cancer.
Clinical Trials
The PIIO is aggressively pursuing IO trials but will focus on investigator-initiated trials, especially the phase I program in testing first-in-man, first-in-class agents, novel IO combinatorial studies and other hypothesis-driven small clinical trials to help answer deep scientific questions.
As a National Cancer Institute (NCI)–designated comprehensive cancer center, the OSUCCC – James offers patients access to many sophisticated cancer clinical trials, including some that involve novel therapies available nowhere else. This means that, when additional treatment options are needed, patients can often find them at the OSUCCC – James.
If you’ve been diagnosed with cancer, would like a second opinion or would like to speak with a cancer specialist about immunotherapy options, please call The James Line at 800-293-5066 or 614-293-5066 to make an appointment.
Immune Monitoring and Discovery Platform (IMDP)
The development of new cancer immunotherapies requires sophisticated assessment of immune responses, development of innovative assays, and possession of advanced technologies and resources. Recent advances in immunophenotyping, spatial imaging, single-cell genomics, and proteomics now facilitate the generation of high dimensional data, the acquisition of abundant information from small amounts of clinical samples, and more accurate and reliable interpretation of data. Human and animal tissue or cells can be measured to monitor changes in immune signatures over time to assess mechanisms of immune function, efficacy of treatments, identification of precision biomarkers and progression/remission of cancer. Therefore, the Ohio State University Comprehensive Cancer Center - Arthur G. James Cancer Hospital and Richard J. Solove Research Institute (OSUCCC-James) Pelotonia Institute for Immuno-Oncology (PIIO) has developed the Immune Monitoring and Discovery Platform (IMDP) to leverage these new advances in technology to provide comprehensive cell- and molecule-based immunoassay services to support basic, translational, and clinical immuno-oncology (IO) studies. The IMDP is no standard shared resource core. Rather, it operates as a technological hub for innovative IO research, paving the way for advanced immune phenotyping and functional analyses as well as multiplexed biomarker detection discovery methods.
The platform’s mission is to mix state-of-the-art instrumentation, high levels of expertise, and exceptional customer service to create an environment that fosters creativity, collaboration, and productivity (Figure 1). The IMDP delivers high-content spectral flow cytometry and cell sorting, mass cytometry, highly-multiplexed tissue imaging, monoclonal antibody production and purification, single cell proteomics and genomics services as well as related accessory equipment with an emphasis on automation. The platform offers QA/QC for all instrumentation, training, experimental design, troubleshooting, and general assistance to users for all services, from the point of experimentation to publication and/or grant application.
The IMDP has five specific aims: 1) Provide cutting-edge IO focused technology that gives researchers a panoramic view of the immune system with regard to cancer research and treatment; 2) In concert with the PIIO’s Immuno-Informatics Group, provide data analytics for flow cytometry, CyTOF, scRNA-seq, scATAC-seq, single-cell proteomics and genomics, bulk RNA-seq, ChIP-seq, ATAC-seq, cytokine data, and spatial imaging; 3) With experts in antibody and protein production, generate novel and high-quality immune reagents, including therapeutic antibodies and recombinant fusion proteins, which will facilitate development of next generation IO biologicals; 4) Develop and maintain an IO Bank as a comprehensive platform for collecting and preserving fresh cells and tissue from IO trial patients and routine clinical care patients, all linked to clinical and research data in real time; and 5) Train and mentor investigators on advanced immune phenotyping and multiplex technologies and novel immunoassay reagent generation.
Future Vision
The IMDP is currently equipped with the latest instrumentation available. The platform conducts experiments using established best practices and optimized standard operating procedures. With this foundation, the IMDP will drive innovation to establish new and better techniques and technologies to push the field forward as a leader in immuno-oncology research. The IMDP is already involved in multiple early-adopter agreements and beta-testing with industry partners to ensure we bring the newest and best technology to the OSUCCC first. As it strengthens its position and further solidifies its presence, the IMDP will engage significantly with external business partners and facilitate collaboration between external biotech and pharmaceutical companies with PIIO scientists and clinicians.
Current Equipment
System 1 – Flow Cytometer: Cytek Aurora Spectral Analyzer, 5 lasers
The Aurora from Cytek Biosciences represents the state-of-the-art in the field of flow cytometry using spectral deconvolution technology to measure large numbers of fluorescent reporters. This instrument is configured with 5 lasers (355nm, 405nm, 488nm, 561nm and 633nm) and 64 detectors to give OSUCCC members tremendous phenotyping depth and maximum flexibility for cancer research. The Aurora can be used to measure live or fixed cells or subcellular particles with a high rate of acquisition (>25K cells/second). While this technology is new it utilizes traditional fluorescent reagents compatible with older technology which allows OSUCCC investigators to rapidly adapt their existing phenotyping panels and then increase the total number of parameters measured as needed, up to 40-color experiments have been run on this system thus far with new reagents rapidly being developed for future expansion.
System 2 – Mass Cytometer: Fluidigm Helios mass cytometer
The Helios mass cytometer from Fluidigm is the third-generation instrument using time-of-flight mass spectrometry for highly parametric single cells assays using metal isotope tagged reagents at single cell resolution. The metal isotope reagents employed on the Helios provide good sensitivity with very minimal overlap in signals when compared with fluorescent technologies and currently allows for the maximum number of parameters for flow cytometry assays in cancer research (up to 59 individual markers have been used). The Helios can be used with fixed cells which can be acquired at a rate of approximately 1K cells/second for optimal sensitivity. The metal tagged reagents are also more robust which allow master mixes of reagents to be frozen for less inter-assay variation over time.
System 3- Flow Cytometer: Cytek Northern Lights spectral analyzer with 3 lasers
Cytek Northern Lights is the only full spectrum unmixing flow cytometer with 3 lasers, 38 FL APD detectors, vacuum fluidics, small particle detection using Violet laser SSC, and capable of 27+ colors. The Northern Lights is a game changer in performance yet the easiest system to learn and teach since no filters are ever needed, compensation is very simplified, and the software is very user friendly.
System 4 – Cell Sorter: BD FACSMelody 3 lasers, 4-way sorting
The FACS Melody cell sorter is a simple, easy to use sorting instrument capable of sorting 4 ways with up to 9 fluorescent colors in to tubes or plates. This instrument features automated setup and optimization for ease of use for the operator. This cell sorter is configured and available for investigators to use whenever convenient with total independence at any time.
System 5 –Spectral cell sorter: Cytek CS 5 lasers
Currently Under Development: 64 channel, 6-way cell sorter, allows sorting of highly parametric flow cytometry samples that has been previously impossible. Estimated availability for testing March 2021.
System 6 – Multiplexed Imaging: Akoya Vectra Polaris
State of the art multispectral imaging enables the identification and downstream quantification of multiple overlapping biomarkers (up to 8) without the interference of autofluorescence as the signals are unmixed from one another. The Vectra Polaris has a renowned liquid crystal tunable filter (LCTF) technology generates unmixed, annotated regions of interest of up to 9-colors for deeper interrogation of biology. MOTiF™ technology generates unmixed whole slide scans at 40x magnification of up to 7 colors in about 20 minutes (15 x 15 mm region) for analysis of biology across the entire slide in a streamlined workflow and without selection bias. Fully automated, high-throughput system to accelerate your research and maximize resources. This system uses integrated inForm and phenoptr tissue analysis software packages support configurable projects for biomarker quantification and spatial analysis
System 7 -Fluorescent microscope with live cell onstage incubator: EVOS 7000
The EVOS M7000 Imaging System offers outstanding image quality and versatility with a 5-position objective turret, 4-color LED fluorescence and transmitted light channels, and 3.2 MP CMOS color and B/W cameras. The EVOS 7000 provides exceptional usability with fully automated X/Y scanning stage, autofocus, and acquisition routines. High-speed image acquisition coupled with multi-position well scanning, and Z-stack and tile-stitch options lend power and flexibility to IO research projects. This system has a fully integrated time-lapse live cell imaging using the EVOS Onstage Incubator for precise control of temperature, gases for normoxic or hypoxic conditions, and humidity
System 8 – Single Cell Epigentics: 10X Genomics Chromium
The Chromium System, using their proprietary Next GEM Technology, provides precisely engineered reagent delivery method that enables thousands of micro-reactions in parallel to measure genomic expression of RNA from single cells. Cell samples are encapsulated into hundreds to tens of thousands of uniquely addressable partitions in minutes, each containing an identifying barcode for downstream analysis. Each cell is captured within a Gel Bead, infused with millions of barcoded oligonucleotides, is mixed with a sample, which can be high molecular weight (HMW) DNA, individual cells, nuclei, or Cell Beads. Gel Beads and samples are then added to an oil-surfactant solution to create Gel Beads-in-Emulsion (GEMs), which act as individual reaction vesicles in which the Gel Beads are dissolved, and the sample is barcoded. Barcoded products are pooled for downstream reactions to create short-read sequencer compatible libraries. After sequencing, the resulting barcoded short read sequences are fed into turnkey analysis pipelines that use the barcode information to map reads back to their original HMW DNA, single cell, or single nucleus of origin.
System 9 – Single Cell Proteomics: Isoplexis IsoLight
The IsoLight is a versatile, precision engineering platform designed to understand and characterize differences among single cells, mapping thousands of cells per sample, to reveal full functional profiles and polyfunctionality among cell subsets to determine patient response and product quality. Running the consumable IsoCode Chip, (known formerly in the literature as single-cell barcode chip or SCBC), the IsoLight system captures singlecell, secretomic, cytokine profiles from thousands of single cells in parallel to better understand complex immunotherapy patient response. The IsoSpeak Informatics platform helps discover new patient relationships amongst heterogeneous cells and clearly defines subsets of powerful, multi-functional protein secreting cells that can help predict patient outcome and determine disease progression. System 10-Automated Electrophoresis: Agilent 4150 TapeStation
The Agilent TapeStation system is an automated electrophoresis solution for the sample quality control of DNA and RNA samples. The system integrates an instrument, data processing software, reagents, and ScreenTape devices specific for DNA and RNA. It is suitable to analyze size, quantity, and integrity of your samples. It fits for example in a next-generation sequencing (NGS) or biobanking workflow with low to high throughput delivering highly precise analytical evaluation.
System 11 – Automated Slide Stainer: Leica BondRx
The BOND RX allow IO researchers fully automate emerging technologies (including new molecular applications), customize all protocol segments and put theories into action. Preparation can be customized to optimize baking and dewax options. Antigen retrieval incubation time and temperature can be altered in any way needed. Staining can be controlled for incubation times, temperatures, dispense types and more.
System 12- Automated Fluid Handler: Eppendorf 5070
The epMotion 5070’s pipetting technology is based on the classic Eppendorf piston-stroke pipettes; thus, protocols previously carried out manually are easily transferred to the liquid handling robot epMotion 5070. Its compatibility to a wide range of predefined consumables also allows established procedures to be maintained. The Eppendorf EasyCon tablet is optimized for simple and comfortable access to liquid handling automation. This automated system is used to complete repetitive processes for IO experiments that are time consuming and sources of possible human error therefor freeing up time for more important work and reducing costly errors that can occur from manual pipetting.
Immuno-Informatics and Data Analytics
Flow Cytometry and Mass Cytometry Data Analysis
The flow cytometry data will be conducted using steps including quality control, scaling, gating, subsampling, dimension reduction, visualization, and differential analysis. First, FCS files will be loaded and data will be transformed to arcsinh scale using the Bioconductor package ‘flowCore’. Second, we will implement rigorous quality control (QC) of the data using various visualizations including density plots, multi-dimensional scaling (MDS) plots, and heatmaps. This includes checking similarity between replicates within the same condition, abnormality of marker expressions, and batch effects. Third, cell population identification will be carried out using both automated gating and manual gating. For the automated gating, cells will be clustered using FlowSOM (1), where we will consider to both manually set various k values and automatically identify an optimal k value using an elbow criterion. The FlowSOM results will be further evaluated using a minimum spanning tree of clusters. Alternatively, we will adopt the multivariate t-mixture model (2) to cluster cells based on the normalized multivariate flow cytometry marker expressions. This robust parametric-model-based approach allows for identifying the optimal number of cell clusters using the Bayesian information criterion (BIC) (3). Thus, for each data set, we can choose the optimal number of cell clusters by selecting the model with minimum BIC. Fourth, data will be visualized using the Uniform Manifold Approximation and Projection (UMAP) algorithm (4), where cells are colored according to either marker expressions or cell clusters. The clustering results will be visually inspected using both UMAP and heatmaps of marker expressions, and cell clusters will be annotated accordingly. Finally, we will implement analysis of cell population abundances using a generalized linear mixed model (GLMM), and of marker intensities within each cluster using a linear mixed model (LMM).
Cytokine and Chemokine, ELISA, and Viral Neutralization Assay
Measurements will first be log-transformed and we will then implement analysis of these log-transformed measurements using LMM.
Bulk RNA-Seq Data Analysis
Sequencing read quality will be evaluated using FastQC and subsequently trimmed to remove adaptor contaminant sequences and low-quality bases using cutadapt (5). Reads pairs with either end too short (<25 bps) will be discarded from further analysis. Next, trimmed and filtered reads will be aligned to the reference transcriptome using STAR aligner (6). HTSeq (7) will be used to generate count data for each sample. Differentially expressed (DE) genes between conditions will be identified using DESeq2 (8) and edgeR (9). The final list of DE genes will be determined at the nominal level of false discovery rate (FDR) and fold change, and DE genes obtained from different algorithms will be combined using multiple approaches (e.g., union and intersection). Enrichment analysis of pathways and Gene Ontology (GO) terms will be carried out using Gene Set Enrichment Analysis (GSEA) (10) and the ToppGene Suite (11).
Bulk ChIP-Seq and ATAC-Seq Data Analysis
After stringent quality control, reads will be aligned to the reference genome using Bowtie2 (12). Peaks will be called using MACS2 (13) and MOSAiCS (14). Peaks obtained from different algorithms will be combined using the irreproducible discovery rate (IDR) approach (15). After identifying peaks, changes in ChIP-seq and ATAC-seq signals between conditions will be identified at the peak level using DiffBind (16) and DBChIP (17), while changes will be identified at the gene level using DESeq2 (8) and edgeR (9). HOMER (18) will be used to annotate the peaks and perform the transcription factor (TF) binding motif enrichment.
T Cell Receptor Repertoire Sequencing Data Analysis
We will use the immunarch workflow to analyze the T cell receptor repertoire sequencing (TCR-seq) data. First, sequencing files will be parsed to take the amino acid sequences after removing unproductive sequences. Clones with different nucleotide sequences but the same amino acid sequence will be aggregated together under one amino acid sequence and their reads will be accumulated. Second, we will check various descriptive statistics, including number of clonotypes, distribution of CDR3 lengths and clonotype abundances, top clonality proportion, rare clonal proportion, relative abundance, and V/D/J gene usage. Third, TCR repertoire will be quantified in the sense of diversity, richness, and evenness, e.g., using Gini-Simpson index and inverse Simpson index. Fourth, clonotypes will be annotated using immune receptor databases, e.g., using VDJDB. Finally, we will implement clustering analysis and repertoire overlap analysis among samples, using Morisita’s overlap index applied to clonotypes and gene usage. We will also implement tracking analysis of the most abundant clonotypes across samples and time points.
Single-Cell RNA-Seq Data Analysis
We will use the Cell Ranger-Seurat workflow to analyze the single cell RNA-seq (scRNA-seq) data. First, using the Cell Ranger software, we will convert BCL files into FASTQ files, trim adapter and primer sequences, map reads to the reference genome, and quantify expressions. In this step, to eliminate low-quality and dying cells, we will filter out cells with counts less than 200 and those with >5% mitochondrial counts. Then, we will use the Seurat software for downstream analyses. First, we will normalize counts using the LogNormalize approach, visualize cells using the UMAP and t-SNE algorithms (19), and determine cell clusters using the shared nearest neighbor (SNN) modularity optimization-based clustering algorithm. Second, we will identify cell type markers conserved between conditions. Third, we will annotate each of these cell clusters with cell types using both the automated annotation (e.g., singleR (20)) and the manual annotation based on canonical cell markers (e.g., PanglaoDB (21) and literature). Finally, for each cluster, we will determine DE genes between conditions using a Wilcoxon Rank Sum test and adjust DE p-values for multiple testing using the Benjamini-Hochberg procedure (22) of the nominal level of 0.05. We will also implement various downstream analyses, including pseudotime analysis using monocle (23), pathway analysis, and cell-type-specific regulon analysis using IRIS3 (24).
Single-Cell ATAC-Seq Data Analysis
We will use the Cell Ranger-Seurat workflow to analyze the single cell ATAC-seq (scATAC-seq) data. Briefly, using the Cell Ranger ATAC analysis pipeline, we will convert BCL files into FASTQ files, trim adapter and primer sequences, map reads to the reference genome, quantify chromatin accessibility levels, and cluster differential accessibility. To eliminate low-quality and dying cells, we will filter out cells with counts less than 200 and those with >5% mitochondrial counts. Then, we will use the Cell Ranger ATAC software with the count data and the peak matrix for the downstream analysis. This peak matrix will be identified using a ZINBA-like (25) mixture model consisting of geometric distribution to model zero-inflated count, negative binomial distribution to model noise, and another geometric distribution to model the signal. We will annotate peaks based on genomic regions (e.g., transcription start sites) and implement transcription factor motif enrichment analysis. Finally, in each cluster, we will implement differential accessibility analysis based on a negative binomial regression model and a Wald test.
Spatial Transcriptomics Data Analysis
We will use the SpaceRanger-Seurat workflow to analyze the spatial transcriptomics data. Using Space Ranger, we will first implement file format conversion, mapping reads to the reference genome, and expression quantification. To eliminate low-quality and dying cells, cells with low counts and those with nonnegligible mitochondrial proportions will be filtered out. Then, we will use the Seurat software for downstream analyses. First, we will normalize counts using the LogNormalize approach, and determine cell clusters using the SNN modularity optimization-based clustering algorithm. Second, we will visualize cells using the UMAP and t-SNE algorithms (19), and also overlay cells on the image. Third, we will identify spatially variable genes, i.e., genes of which expressions correlate with spatial location within a tissue, based on a mark point process (26).
Mining Public Data
The Cancer Genome Atlas (TCGA) will be mined using cBioPortal (http://www.cbioportal.org) and the NCI Genomic Data Commons Data Portal (https://portal.gdc.cancer.gov/), including genomic data (e.g., mRNA expression and somatic mutations) and relevant clinical data (overall survival, progress-free survival, sex, race, etc.). It will be further integrated with immunologic parameters from the Immune Landscape of Cancer (27) and its data repository (https://gdc.cancer.gov/about-data/publications/panimmune). The available data includes cellular fraction estimates (leukocyte and stromal fractions, CIBERSORT immune fractions), TCR and B cell receptor repertoire (BCR) richness and evenness, and aneuploidy score. In addition, we will mine other cancer genomic data repositories as appropriately, including the Cancer Proteome Atlas (TCPA; https://www.tcpaportal.org/tcpa/) and ImmGen (https://www.immgen.org/), among others.
Other technologies
Other data analytics tools being developed include single cell proteomics data analytic pipeline, spatial imaging data analysis, neoantigen discovery, HLA repertoire analysis, and artificial intelligence.
Key Personnel:
Zihai Li, MD, PhD, Founding Director, Pelotonia Institute for Immuno-Oncology
Kevin Weller, BS, Associate Director, Immune Monitoring and Discovery Platform (IMDP)
Qin Ma, PhD, Leader of the Immuno-Oncology Informatics Group (IOIG)
Dongjun Chung, PhD, Associate Professor, IOIG
Brian Searle, PhD, Assistant Professor, IOIG, Leader of IO Proteomics
Donna Bucci, BS, IO-Bank Project Manager
Yuzhou Chang, PhD Candidate, IOIG
Komal Das, PhD, Research Scientist, IMDP
Bob Davenport, BA, Research Specialist, IMDP
Jamie Hamon, MS, Senior Manager, IMDP
Zayneb Hussein, BS, Research Assistant, IMDP
Anjun Ma, PhD, Research Scientist, IOIG
Micah Marshall, BS, Research Assistant, IMDP
Kelsi Reynolds, BS, Research Assistant, IMDP
References
- Van Gassen S, Callebaut B, Van Helden MJ, Lambrecht BN, Demeester P, Dhaene T, Saeys Y. FlowSOM: Using self-organizing maps for visualization and interpretation of cytometry data. Cytometry A. 2015;87(7):636-45. Epub 2015/01/13. doi: 10.1002/cyto.a.22625. PubMed PMID: 25573116.
- Lo K, Brinkman RR, Gottardo R. Automated gating of flow cytometry data via robust model‐based clustering. Cytometry Part A: the journal of the International Society for Analytical Cytology. 2008;73(4):321-32.
- Schwarz G. Estimating the dimension of a model. Annals of statistics. 1978;6(2):461-4.
- Becht E, McInnes L, Healy J, Dutertre CA, Kwok IWH, Ng LG, Ginhoux F, Newell EW. Dimensionality reduction for visualizing single-cell data using UMAP. Nat Biotechnol. 2018. Epub 2018/12/12. doi: 10.1038/nbt.4314. PubMed PMID: 30531897.
- M M. Cutadapt Removes Adapter Sequences from High-Throughput Sequencing Reads. EMBnet Journal. 2011;17:10-2. doi: org/10.14806/ej.17.1.200
- Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M, Gingeras TR. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 2013;29(1):15-21. Epub 2012/10/30. doi: 10.1093/bioinformatics/bts635. PubMed PMID: 23104886; PMCID: PMC3530905.
- Anders S, Pyl PT, Huber W. HTSeq--a Python framework to work with high-throughput sequencing data. Bioinformatics. 2015;31(2):166-9. Epub 2014/09/28. doi: 10.1093/bioinformatics/btu638. PubMed PMID: 25260700; PMCID: PMC4287950.
- Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550. Epub 2014/12/18. doi: 10.1186/s13059-014-0550-8. PubMed PMID: 25516281; PMCID: PMC4302049.
- Robinson MD, McCarthy DJ, Smyth GK. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics. 2010;26(1):139-40. Epub 2009/11/17. doi: 10.1093/bioinformatics/btp616. PubMed PMID: 19910308; PMCID: PMC2796818.
- Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102(43):15545-50. Epub 2005/10/04. doi: 10.1073/pnas.0506580102. PubMed PMID: 16199517; PMCID: PMC1239896.
- Chen J, Bardes EE, Aronow BJ, Jegga AG. ToppGene Suite for gene list enrichment analysis and candidate gene prioritization. Nucleic acids research. 2009;37(Web Server issue):W305-11. Epub 2009/05/26. doi: 10.1093/nar/gkp427. PubMed PMID: 19465376; PMCID: PMC2703978.
- Langmead B, Trapnell C, Pop M, Salzberg SL. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol 2009;10(3):R25.
- Zhang Y, Liu T, Meyer CA, Eeckhoute J, Johnson DS, Bernstein BE, Nusbaum C, Myers RM, Brown M, Li W, Liu XS. Model-based analysis of ChIP-Seq (MACS). Genome Biol. 2008;9(9):R137. Epub 2008/09/19. doi: 10.1186/gb-2008-9-9-r137. PubMed PMID: 18798982; PMCID: PMC2592715.
- Kuan PF, Chung D, Pan G, Thomson JA, Stewart R, Keles S. A Statistical Framework for the Analysis of ChIP-Seq Data. J Am Stat Assoc. 2011;106(495):891-903. Epub 2011/01/01. doi: 10.1198/jasa.2011.ap09706. PubMed PMID: 26478641; PMCID: PMC4608541.
- Li Q, Brown JB, Huang H, Bickel PJ. Measuring reproducibility of high-throughput experiments. The annals of applied statistics. 2011;5(3):1752-79.
- Ross-Innes CS, Stark R, Teschendorff AE, Holmes KA, Ali HR, Dunning MJ, Brown GD, Gojis O, Ellis IO, Green AR, Ali S, Chin SF, Palmieri C, Caldas C, Carroll JS. Differential oestrogen receptor binding is associated with clinical outcome in breast cancer. Nature. 2012;481(7381):389-93. Epub 2012/01/06. doi: 10.1038/nature10730. PubMed PMID: 22217937; PMCID: PMC3272464.
- Liang K, Keles S. Detecting differential binding of transcription factors with ChIP-seq. Bioinformatics. 2012;28(1):121-2. Epub 2011/11/08. doi: 10.1093/bioinformatics/btr605. PubMed PMID: 22057161; PMCID: PMC3244766.
- Heinz S, Benner C, Spann N, Bertolino E, Lin YC, Laslo P, Cheng JX, Murre C, Singh H, Glass CK. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol Cell. 2010;38(4):576-89. Epub 2010/06/02. doi: 10.1016/j.molcel.2010.05.004. PubMed PMID: 20513432; PMCID: PMC2898526.
- Maaten LV, Hinton G. Visualizing data using t-SNE. Journal of Machine Learning Research. 2008;9(Nov):2579-605.
- Aran D, Looney AP, Liu L, Wu E, Fong V, Hsu A, Chak S, Naikawadi RP, Wolters PJ, Abate AR, Butte AJ, Bhattacharya M. Reference-based analysis of lung single-cell sequencing reveals a transitional profibrotic macrophage. Nat Immunol. 2019;20(2):163-72. Epub 2019/01/16. doi: 10.1038/s41590-018-0276-y. PubMed PMID: 30643263; PMCID: PMC6340744.
- Franzén O, Gan LM, Björkegren JLM. PanglaoDB: a web server for exploration of mouse and human single-cell RNA sequencing data. Database (Oxford). 2019;2019. Epub 2019/04/06. doi: 10.1093/database/baz046. PubMed PMID: 30951143; PMCID: PMC6450036.
- Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. . Journal of the Royal Statistical Society: Series B (Methodological). 1995 Jan;57(1):289-300.
- Trapnell C, Cacchiarelli D, Grimsby J, Pokharel P, Li S, Morse M, Lennon NJ, Livak KJ, Mikkelsen TS, Rinn JL. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat Biotechnol. 2014;32(4):381-6. Epub 2014/03/25. doi: 10.1038/nbt.2859. PubMed PMID: 24658644; PMCID: PMC4122333.
- Ma A, Wang C, Chang Y, Brennan FH, McDermaid A, Liu B, Zhang C, Popovich PG, Ma Q. IRIS3: integrated cell-type-specific regulon inference server from single-cell RNA-Seq. Nucleic Acids Research. 2020;48(W1):W275-W86. doi: 10.1093/nar/gkaa394; PMCID: PMC7319566.
- Rashid NU, Giresi PG, Ibrahim JG, Sun W, Lieb JD. ZINBA integrates local covariates with DNA-seq data to identify broad and narrow regions of enrichment, even within amplified genomic regions. Genome biology. 2011;12(7):R67. doi: 10.1186/gb-2011-12-7-r67. PubMed PMID: 21787385; PMCID: 3218829.
- Edsgärd D, Johnsson P, Sandberg R. Identification of spatial expression trends in single-cell gene expression data. Nat Methods. 2018;15(5):339-42. Epub 2018/03/20. doi: 10.1038/nmeth.4634. PubMed PMID: 29553578; PMCID: PMC6314435.
- Thorsson V, Gibbs DL, Brown SD, Wolf D, Bortone DS, Ou Yang TH, Porta-Pardo E, Gao GF, Plaisier CL, Eddy JA, Ziv E, Culhane AC, Paull EO, Sivakumar IKA, Gentles AJ, Malhotra R, Farshidfar F, Colaprico A, Parker JS, Mose LE, Vo NS, Liu J, Liu Y, Rader J, Dhankani V, Reynolds SM, Bowlby R, Califano A, Cherniack AD, Anastassiou D, Bedognetti D, Mokrab Y, Newman AM, Rao A, Chen K, Krasnitz A, Hu H, Malta TM, Noushmehr H, Pedamallu CS, Bullman S, Ojesina AI, Lamb A, Zhou W, Shen H, Choueiri TK, Weinstein JN, Guinney J, Saltz J, Holt RA, Rabkin CS, Cancer Genome Atlas Research N, Lazar AJ, Serody JS, Demicco EG, Disis ML, Vincent BG, Shmulevich I. The Immune Landscape of Cancer. Immunity. 2018;48(4):812-30 e14. Epub 2018/04/10. doi: 10.1016/j.immuni.2018.03.023. PubMed PMID: 29628290; PMCID: PMC5982584.