Identification Of Subset Research Articles

P1.04-35 Identification and Characterization of a Unique KLRG1-Expressing Subset of CD4+FOXP3+ Tregs in Non-Small Cell Lung Cancer

It is well established that CD4+FOXP3+ T regulatory cells (Tregs) contribute to dampening anti-tumor responses. Despite our improved understanding of their role in cancer, there remains a knowledge gap with respect to the clonal composition and heterogeneity of tumor-infiltrating Tregs in solid cancers. We addressed this outstanding issue by conducting comprehensive phenotypic profiling of Tregs present in lung adenocarcinomas of a genetically engineered mouse model of non-small cell lung cancer (NSCLC) as well as those in resected tumors of NSCLC patients. Multi-parameter flow cytometric analysis revealed that unlike the peripheral tissues, the tumor harbors a distinct sub-population of Tregs that express the co-inhibitory receptor, KLRG1. Compared to their negative counterparts, the KLRG1+ Treg subset exhibited heightened expression of a number of Treg signature proteins as well as higher levels of activation and memory molecular markers suggesting that they are a highly activated and differentiated Treg pool that is recruited to, or induced in the tumor microenvironment. Consistent with this phenotype, these KLRG1+ Tregs were superior in their capacity to suppress T cell proliferation relative to the KLRG1- cells. Collectively, these findings demonstrate that the tumor microenvironment in non-small cell lung cancer harbors a unique Treg sub-population that is characterized by dominant expression of KLRG1, and which represents a clonal pool with the most potent inhibition of T cell responses. These studies highlight the installment of distinct Treg subsets in NSCLC that have implications for regulation of anti-tumor responses.

Identification of an excellent prognosis subset of human papillomavirus-associated oropharyngeal cancer patients by quantification of intratumoral CD103+ immune cell abundance.

Accurate prognostic stratification of human papillomavirus-associated oropharyngeal cancers (HPV+OPSCC) is required to identify patients potentially suitable for treatment deintensification. We evaluated the prognostic significance of CD103, a surface marker associated with tissue-resident memory T cells (TRMs), in two independent cohorts of patients with HPV+OPSCC. The abundance and distribution of CD103+ immune cells were quantified using immunohistochemistry in a cohort of 189 HPV+OPSCC patients treated with curative intent and correlated with outcome. Findings were then validated in an independent cohort comprising 177 HPV+OPSCCs using univariable and multivariable analysis. Intratumoral CD103+ immune cells were characterized by multispectral fluorescence immunohistochemistry and gene expression analysis. High intratumoral abundance of CD103+ immune cells using a ≥30% cut-off was found in 19.8% of tumors in the training cohort of HPV+OPSCC patients and associated with excellent prognosis for overall survival (OS) with adjusted hazard ratio (HR) of 0.13 [95% confidence interval (CI) 0.02-0.94, P = 0.004]. In the independent cohort of HPV+OPSCCs, 20.4% had high intratumoral CD103+ abundance and an adjusted HR for OS of 0.16 (95% CI 0.02-1.22, P = 0.02). Five year OS of patients with high intratumoral CD103 was 100% across both cohorts. The C-statistic for the multivariate prognostic model with stage and age was significantly improved in both cohorts with the addition of intratumoral CD103+ cell abundance. On the basis of spatial location, co-expression of CD8 and CD69, and gene expression profiles, intratumoral CD103+ cells were consistent with TRMs. Quantification of intratumoral CD103+ immune cell abundance provides prognostic information beyond that provided by clinical parameters such as TNM-staging, identifying a population of low risk HPV+OPSCC patients who are good candidates for trials of deintensification strategies.

Dissection of gastric cancer heterogeneity for precision oncology.

Gastric cancer (GC) remains the fifth most prevalent cancer worldwide and the third leading cause of global cancer mortality. Comprehensive ‐omic studies have unveiled a heterogeneous GC landscape, with considerable molecular diversity both between and within tumors. Given the complex nature of GC, a long‐sought goal includes effective identification of distinct patient subsets with prognostic and/or predictive outcomes to enable tailoring of specific treatments (“precision oncology”). In this review, we highlight various approaches to molecular classification in GC, covering recent genomic, transcriptomic, proteomic and epigenomic features. We pay special attention to the translational significance of classifier systems and examine potential confounding factors which deserve further investigation. In particular, we discuss recent advancements in our knowledge of intra‐subtype, intra‐patient and intra‐tumor heterogeneity, and the pivotal role of the tumor stromal microenvironment.

IL-13 mRNA Tissue Content Identifies Two Subsets of Adult Ulcerative Colitis Patients With Different Clinical and Mucosa-Associated Microbiota Profiles.

A personalized approach to therapy hold great promise to improve disease outcomes. To this end, the identification of different subsets of patients according to the prevalent pathogenic process might guide the choice of therapeutic strategy. We hypothesize that ulcerative colitis [UC] patients might be stratified according to distinctive cytokine profiles and/or to a specific mucosa-associated microbiota. In a cohort of clinically and endoscopic active UC patients and controls, we used quantitative PCR to analyse the mucosal cytokine mRNA content and 16S rRNA gene sequencing to assess the mucosa-associated microbiota composition. We demonstrate, by means of data-driven approach, the existence of a specific UC patient subgroup characterized by elevated IL-13 mRNA tissue content separate from patients with low IL-13 mRNA tissue content. The two subsets differ in clinical-pathological characteristics. High IL-13 mRNA patients are younger at diagnosis and have a higher prevalence of extensive colitis than low IL-13 mRNA patients. They also show more frequent use of steroid/immunosuppressant/anti-tumour necrosis factor α therapy during 1 year of follow-up. The two subgroups show differential enrichment of mucosa-associated microbiota genera with a prevalence of Prevotella in patients with high IL-13 mRNA tissue content and Sutterella and Acidaminococcus in patients with low IL-13 mRNA tissue content. Assessment of mucosal IL-13 mRNA might help in the identification of a patient subgroup that might benefit from a therapeutic approach modulating IL-13. This article has an associated podcast which can be accessed at https://academic.oup.com/ecco-jcc/pages/podcast.

Evaluation of a 10color protocol as part of a 2tube screening panel for flow cytometric assessment of peripheral blood leukocytic subsets

Peripheral blood (PB) immunophenotyping is commonly required for initial evaluation of various suspected disease entities. Several approaches have been proposed. The objective of this work is to explore the value of a 10color protocol developed in our laboratory for flow cytometric assessment of PB leukocytic subsets, as part of a 2tube screening panel. A combination of CD16/CD56/CD34/CD33/CD19/CD4/CD8/CD3/CD20/CD45 antibodies in 1 tube was applied routinely during flow cytometric analysis of PB samples for diagnostic purposes. The protocol was systematically complemented by a 2nd tube with anti-kappa, anti-lambda, CD5, CD19, and CD45 antibodies for adults and selected pediatric patients, and specifically oriented panels when necessary. 25 samples with no detectable neoplastic PB involvement and 31 samples with a hematolymphoid disorder were investigated retrospectively. The contribution of CD33 in the separation of leukocytic populations, as well as the benefits from the simultaneous assessment of CD20/CD19/CD45, CD16/CD56 and the detection of CD34+ cells were examined. The gating strategy with the use of CD33 provided additional information in certain cases. The protocol enabled recognition of differential expression of CD20 and CD45 in CD19+ cells with chronic lymphocytic leukemia phenotype, overall evaluation of NK and NK like T cells, estimation of CD16− granulocytes and CD56/CD16 expression in monocytes, as well as identification of minor cell subsets, such as CD34+ cells. The proposed 10color combination of antibodies analyzed in a standardized manner can offer significant information in the initial evaluation of PB samples, thus, guiding subsequent investigation if needed.

P14.48 Identification of subsets of IDH-mutant glioblastomas with distinct epigenetic and copynumber alterations and stratified clinical risks

Abstract BACKGROUND IDH-mutant glioblastoma is classified by the 2016 CNS WHO as a group with good prognosis. However, the actual number of cases examined in the literature is relatively small. We hypothesize that IDH-mutant glioblastoma is not a uniform group and should be further stratified. MATERIAL AND METHODS We conducted methylation profiles and estimated copy number variations in 64 IDH-mutant glioblastomas. RESULTS Our results showed that 10.9%, 53.1%, and 35.9% of tumors belonged to Codel, G-CIMPhigh, and G-CIMP-low methylation subgroups, respectively. G-CIMP-low subgroup was associated with significantly worse OS as compared to G-CIMP-high (P=0.005) and Codel groups (P=0.009). CDKN2A deletion (37.5%) was the most common gene copy number variation, and was associated with G-CIMP-low subgroup (P=0.001). Other frequent copy number changes included MET (4.7%), CCND2 (17.2%), PDGFRA (14.1%), CDK4 (12.5%), and EGFR (12.5%) amplification. Both CDKN2A deletion (P=0.008) and MET amplification (P<0.001) were associated with poor OS in IDH-mutant glioblastomas. Combined epigenetic signatures and gene copy number variations separated IDH-mutant glioblastomas into Group 1 (Codel), Group 2 (GCIMP- high), Group 3 (G-CIMP-low without CDKN2A and MET alterations), and Group 4 (G-CIMP-low with CDKN2A/MET alteration). Survival analysis revealed Group 1 had a favorable OS (median survival: 41.3 months), while Groups 2 and 3 exhibited an intermediate OS (median survival: 20.6 and 21.8 months, respectively). Group 4 exhibited the worst OS (median survival: 8.4 months). Multivariable analysis confirmed the independent prognostic significance of our Groups. CONCLUSION IDH-mutant glioblastomas should be stratified for risk with combined epigenetic signature and CDKN2A/MET status and some cases have poor outcome.

Single-Cell Analysis of Human Mononuclear Phagocytes Reveals Subset-Defining Markers and Identifies Circulating Inflammatory Dendritic Cells

Human mononuclear phagocytes comprise phenotypically and functionally overlapping subsets of dendritic cells (DCs) and monocytes, but the extent of their heterogeneity and distinct markers for subset identification remains elusive. By integrating high-dimensional single-cell protein and RNA expression data, we identified distinct markers todelineate monocytes from conventional DC2 (cDC2s). Using CD88 and CD89 for monocytes and HLA-DQ and FcεRIα for cDC2s allowed for their specific identification in blood and tissues. We also showed that cDC2s could be subdivided into phenotypically and functionally distinct subsets based on CD5, CD163, and CD14 expression, including a distinct subset of circulating inflammatory CD5-CD163+CD14+ cells related to previously defined DC3s. These inflammatory DC3s were expanded in systemic lupus erythematosus patients and correlated with disease activity. These findings further unravel the heterogeneity of DC subpopulations in health and disease and may pave the way for the identification of specific DC subset-targeting therapies.

EuroFlow Lymphoid Screening Tube (LST) data base for automated identification of blood lymphocyte subsets

In recent years the volume and complexity of flow cytometry data has increased substantially. This has led to a greater number of identifiable cell populations in a single measurement. Consequently, new gating strategies and new approaches for cell population definition are required. Here we describe how the EuroFlow Lymphoid Screening Tube (LST) reference data base for peripheral blood (PB) samples was designed, constructed and validated for automated gating of the distinct lymphoid (and myeloid) subsets in PB of patients with chronic lymphoproliferative disorders (CLPD). A total of 46 healthy/reactive PB samples which fulfilled pre-defined technical requirements, were used to construct the LST-PB reference data base. In addition, another set of 92 PB samples (corresponding to 10 healthy subjects, 51 B-cell CLPD and 31 T/NK-cell CLPD patients), were used to validate the automated gating and cell-population labeling tools with the Infinicyt software.An overall high performance of the LST-PB data base was observed with a median percentage of alarmed cellular events of 0.8% in 10 healthy donor samples and of 44.4% in CLPD data files containing 49.8% (range: 1.3–96%) tumor cells. The higher percent of alarmed cellular events in every CLPD sample was due to aberrant phenotypes (75.6% cases) and/or to abnormally increased cell counts (86.6% samples). All 18 (22%) data files that only displayed numerical alterations, corresponded to T/NK-cell CLPD cases which showed a lower incidence of aberrant phenotypes (41%) vs B-cell CLPD cases (100%). Comparison between automated vs expert-bases manual classification of normal (r2 = 0.96) and tumor cell populations (rho = 0.99) showed a high degree of correlation.In summary, our results show that automated gating of cell populations based on the EuroFlow LST-PB data base provides an innovative, reliable and reproducible tool for fast and simplified identification of normal vs pathological B and T/NK lymphocytes in PB of CLPD patients.

Separation of Low- Versus High-grade Crohn's Disease-associated Small Bowel Carcinomas is Improved by Invasive Front Prognostic Marker Analysis.

Crohn's disease-associated small bowel carcinoma is a rare event, usually reported to have a severe prognosis. However, in previous investigations we have found a minority of cases displaying a relatively favourable behaviour, thus outlining the need to improve the histopathological prediction of Crohn's disease-associated small bowel carcinoma prognosis. As in recent studies on colorectal cancer, a substantial improvement in prognostic evaluations has been provided by the histological analysis of the tumour invasive front; we therefore systematically analysed the tumour budding and poorly differentiated clusters in the invasive front of 47 Crohn's disease-associated small bowel carcinomas collected through the Small Bowel Cancer Italian Consortium. Both tumour budding and poorly differentiated cluster analyses proved highly effective in prognostic evaluation of Crohn's disease-associated small bowel carcinomas. In addition, they retained prognostic value when combined with two other parameters, i.e. glandular histology and stage I/II, both known to predict a relatively favourable small bowel carcinoma behaviour. In particular, association of tumour budding and poorly differentiated clusters in a combined invasive front score allowed identification of a minor subset of cancers [12/47, 25%] characterised by combined invasive front low grade coupled with a glandular histology and a low stage [I or II] and showing no cancer-related death during a median follow-up of 73.5 months. The improved distinction of lower- from higher-grade Crohn's disease-associated small bowel carcinomas provided by invasive front analysis should be of potential help in choosing appropriate therapy for these rare and frequently ominous neoplasms.

Comprehensive Cell Surface Antigen Analysis Identifies Transferrin Receptor Protein-1 (CD71) as a Negative Selection Marker for Human Neuronal Cells.

Cell state‐, developmental stage‐, and lineage‐specific combinatorial expression of cluster of differentiation (CD) molecules enables the identification of cellular subsets via multicolor flow cytometry. We describe an exhaustive characterization of neural cell types by surface antigens, exploiting human pluripotent stem cell‐derived neural cell systems. Using multiwell screening approaches followed by detailed validation of expression patterns and dynamics, we exemplify a strategy for resolving cellular heterogeneity in stem cell paradigms. In addition to providing a catalog of surface antigens expressed in the neural lineage, we identified the transferrin receptor‐1 (CD71) to be differentially expressed in neural stem cells and differentiated neurons. In this context, we describe a role for N‐Myc proto‐oncogene (MYCN) in maintaining CD71 expression in proliferating neural cells. We report that in vitro human stem cell‐derived neurons lack CD71 surface expression and that the observed differential expression can be used to identify and enrich CD71− neuronal derivatives from heterogeneous cultures. stem cells 2019;37:1293–1306

VirtualCytometry: a webserver for evaluating immune cell differentiation using single-cell RNA sequencing data.

The immune system has diverse types of cells that are differentiated or activated via various signaling pathways and transcriptional regulation upon challenging conditions. Immunophenotyping by flow and mass cytometry are the major approaches for identifying key signaling molecules and transcription factors directing the transition between the functional states of immune cells. However, few proteins can be evaluated by flow cytometry in a single experiment, preventing researchers from obtaining a comprehensive picture of the molecular programs involved in immune cell differentiation. Recent advances in single-cell RNA sequencing (scRNA-seq) have enabled unbiased genome-wide quantification of gene expression in individual cells on a large scale, providing a new and versatile analytical pipeline for studying immune cell differentiation. We present VirtualCytometry, a web-based computational pipeline for evaluating immune cell differentiation by exploiting cell-to-cell variation in gene expression with scRNA-seq data. Differentiating cells often show a continuous spectrum of cellular states rather than distinct populations. VirtualCytometry enables the identification of cellular subsets for different functional states of differentiation based on the expression of marker genes. Case studies have highlighted the usefulness of this subset analysis strategy for discovering signaling molecules and transcription factors for human T-cell exhaustion, a state of T-cell dysfunction, in tumor and mouse dendritic cells activated by pathogens. With more than 226 scRNA-seq datasets precompiled from public repositories covering diverse mouse and human immune cell types in normal and disease tissues, VirtualCytometry is a useful resource for the molecular dissection of immune cell differentiation. www.grnpedia.org/cytometry.

Identification of a distinct subset of disease-associated gain-of-function missense mutations in the STAT1 coiled-coil domain as system mutants

Heterozygous gain-of-function (GOF) mutations in the cytokine-regulated transcription factor STAT1 (signal transducer and activator of transcription 1) lead to chronic mucocutaneous candidiasis (CMC). However, the molecular basis of these pathogenic missense mutations is largely unknown. In this study, we characterized in more detail the CMC-associated GOF substitution mutation of arginine-to-tryptophan at position 274 (R274W) and, in addition, the adjacent glutamine-to-alanine mutation at position 275 (Q275A). Both mutants displayed elevated tyrosine phosphorylation levels, prolonged nuclear accumulation, and increased transcriptional responses to interferon-γ (IFNγ) stimulation. No difference was observed between wild-type (WT) and mutant STAT1 in DNA sequence-specificity or dissociation kinetics from high-affinity DNA-binding elements known as gamma-activated sites (GAS). Furthermore, all variants exhibited similar cooperative DNA binding. Unexpectedly, in vitro dephosphorylation rates using the recombinant STAT1-inactivating Tc45 phosphatase in both the absence and presence of double-stranded GAS elements were similar in all STAT1 variants. Likewise, the rate of tyrosine phosphorylation by Janus kinase 2 (JAK2) was unaltered as compared to the WT molecule, excluding that the phenotype of these mutants is caused by either defective Tc45-catalyzed dephosphorylation or JAK2-induced hyper-activation. Interestingly, within 10 min of IFNγ exposure, the majority of R274W and Q275A molecules had entered the nucleus, whereas the wild-type protein remained predominantly cytosolic. Thus, the exchange of critical residues located at the binding interface in the antiparallel dimer conformer led to a premature accumulation of phospho-STAT1 in the nuclear compartment. In summary, our data show that the hyper-activity of the GOF mutations results, at least in part, from the premature nuclear import of the tyrosine-phosphorylated molecules and not from alterations in their phosphorylation or dephosphorylation rates.

Hitting back at lymphoma: How do modern diagnostics identify high-risk diffuse large B-cell lymphoma subsets and alter treatment?

Diffuse large B-cell lymphoma (DLBCL) is a clinically and biologically heterogeneous disease. Diagnostic tools in the clinic can now identify distinct subsets characterized by unique molecular features, which are increasingly transforming how these patients are managed. Activated B-cell-like DLBCL is characterized by NF-κB activation and chronic B-cell receptor signaling and may be targeted with lenalidomide or ibrutinib in the relapsed setting. Germinal center-like DLBCL is enriched for activating EZH2 mutations, and encouraging activity has been observed for the EZH2 inhibitor tazemetostat, which now has a fast-track US Food and Drug Administration designation. Double-hit lymphoma is a high-grade B-cell lymphoma characterized by translocations of MYC and BCL2 and/or BCL6 and carries a poor prognosis. Intensive chemoimmunotherapy strategies appear to be superior to standard R-CHOP (rituximab combined with cyclophosphamide, doxorubicin, vincristine, and prednisone) as initial therapy, and anti-CD19 chimeric antigen receptor T cells are inducing remission in patients with relapsed/refractory disease who previously had few available options. Primary mediastinal (thymic) large B-cell lymphoma is a molecularly distinct large-cell lymphoma with clinical and molecular features that overlap with those of classical Hodgkin lymphoma. R-CHOP has been associated with an unacceptably high rate of primary treatment failure in this young population, whereas dose-adjusted EPOCH-R (etoposide, prednisone, vincristine, cyclophosphamide, and doxorubicin plus rituximab) produces durable remissions without the need for radiotherapy in most patients. For relapsed/refractory disease, immune checkpoint inhibitors targeting PD-1 have shown promising activity in chemotherapy-refractory disease, as have anti-CD19 chimeric antigen receptor T cells. Additional therapeutic targets, including JAK2, continue to be evaluated. The identification of discrete biological subsets is steadily moving us away from a "one-size-fits-all" approach in DLBCL.

Abstract LB-319: Establishment of automated multiplex immunofluorescence T-cell panel for baseline screening of syngeneic mouse models and evaluation of T-cell infiltration in A20 syngeneic tumors treated with TAK-981, a first-in-class SUMO inhibitor

Abstract Mouse syngeneic tumor models play a crucial role in the translationally oriented process of assessing the preclinical activity and mechanism of action of emerging novel immunotherapies and immunotherapy combinations. Recent advancements in fluorescent multiplex technology provide a potential opportunity to improve the evaluation and identification of specific subsets of mouse immune cells using a panel approach. This technology may provide the ability to complement more traditional means of immunophenotyping such as flow cytometry. We established Opal™ multiplex immunofluorescence (IF) assays for the evaluation of T-cells (CD4, CD8, CD25, FOXP3 and tumor marker) in mouse using a Leica Bond RX autostainer. Images were acquired by Vectra® Polaris™ multispectral scanner. HALO™ image analysis was used to assess identification, quantification, and spatial localization of the immune cell subsets within tumors. The multiplex IF assays were used to carry out baseline characterization of the T-cell panels in multiple syngeneic models (A20, CT26, 4T1, EMT6, B16F10, LLC1, Renca and MC38). Notably, integrated data from the Opal™ multiplex IF assay enabled characterization and validation of “hot vs. cold” tumors across multiple syngeneic models. In addition, using PAX5 as a tumor marker, the Opal™ multiplex IF assay T-cell panel was used to specifically evaluate A20 lymphoma tumors from BALB/c mice treated with TAK-981, an investigational, first-in-class small ubiquitin-like modifier (SUMO) inhibitor. Currently, TAK-981 is being evaluated in a first-in-human, Phase I trial (NCT03648372) in patients with relapsed/refractory lymphomas. In pre-clinical models, A20 tumors demonstrated a dose-dependent increase in cytotoxic T cell numbers in response to TAK-981, consistent with previously generated flow cytometry data in animal models and indicative of induction of an adaptive anti-tumor immune response by TAK-981. In conclusion, across a range of syngeneic mouse models, an Opal™ multiplex IF assay can successfully enable characterization of “hot vs. cold” tumors. In tumors derived from a specific syngeneic mouse model treated with a novel, first-in-class SUMO inhibitor, the Opal™ multiplex IF assay was also capable of recapitulating the characterization of the T-cell repertoire within the tumor microenvironment previously determined by flow cytometry. Together, this suggests multiplex IF can contribute to ongoing efforts driving early translational research in immuno-oncology. Citation Format: Omid Ghasemi, Agatha Zawadzka, Evan Luongo, Aimy Tse, Johara Chouitar, Cierra Casson, Michelle Ganno-Sherwood, Xingyue He, Patrick LeRoy, Kristina Xega, Stephan Grossman, Keli Song, Allison Berger, Katherine Galvin, Dennis Huszar, Vaishali Shinde. Establishment of automated multiplex immunofluorescence T-cell panel for baseline screening of syngeneic mouse models and evaluation of T-cell infiltration in A20 syngeneic tumors treated with TAK-981, a first-in-class SUMO inhibitor [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr LB-319.

Empirical comparison of the adjustable spanner and the adaptive toolbox models of choice.

Past research indicates that individuals respond adaptively to contextual factors in multiattribute choice tasks. Yet it remains unclear how this adaptation is cognitively governed. In this article, empirically testable implementations of two prominent competing theoretical frameworks are developed and compared across two multiattribute choice experiments: the adaptive toolbox framework assuming discrete choice strategies and the adjustable spanner framework assuming one comprehensive adaptive strategy. Results from two experiments indicate that in the environments we tested, in which all cue information was presented openly, the toolbox makes better predictions than the adjustable spanner both in- and out-of-sample. Follow-up simulation studies indicate that it is difficult to discriminate the models based on choice outcomes alone but allowed the identification of a small subset of cases where the predictions of both models diverged. Our results suggest that people adapt their decision strategies by flexibly switching between using as little information as possible and use of all of the available information. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

Machine learning to predict microbial community functions: An analysis of dissolved organic carbon from litter decomposition.

Microbial communities are ubiquitous and often influence macroscopic properties of the ecosystems they inhabit. However, deciphering the functional relationship between specific microbes and ecosystem properties is an ongoing challenge owing to the complexity of the communities. This challenge can be addressed, in part, by integrating the advances in DNA sequencing technology with computational approaches like machine learning. Although machine learning techniques have been applied to microbiome data, use of these techniques remains rare, and user-friendly platforms to implement such techniques are not widely available. We developed a tool that implements neural network and random forest models to perform regression and feature selection tasks on microbiome data. In this study, we applied the tool to analyze soil microbiome (16S rRNA gene profiles) and dissolved organic carbon (DOC) data from a 44-day plant litter decomposition experiment. The microbiome data includes 1709 total bacterial operational taxonomic units (OTU) from 300+ microcosms. Regression analysis of predicted and actual DOC for a held-out test set of 51 samples yield Pearson’s correlation coefficients of.636 and.676 for neural network and random forest approaches, respectively. Important taxa identified by the machine learning techniques are compared to results from a standard tool (indicator species analysis) widely used by microbial ecologists. Of 1709 bacterial taxa, indicator species analysis identified 285 taxa as significant determinants of DOC concentration. Of the top 285 ranked features determined by machine learning methods, a subset of 86 taxa are common to all feature selection techniques. Using this subset of features, prediction results for random permutations of the data set are at least equally accurate compared to predictions determined using the entire feature set. Our results suggest that integration of multiple methods can aid identification of a robust subset of taxa within complex communities that may drive specific functional outcomes of interest.

Automated subset identification and characterization pipeline for multidimensional flow and mass cytometry data clustering and visualization

When examining datasets of any dimensionality, researchers frequently aim to identify individual subsets (clusters) of objects within the dataset. The ubiquity of multidimensional data has motivated the replacement of user-guided clustering with fully automated clustering. The fully automated methods are designed to make clustering more accurate, standardized and faster. However, the adoption of these methods is still limited by the lack of intuitive visualization and cluster matching methods that would allow users to readily interpret fully automatically generated clusters. To address these issues, we developed a fully automated subset identification and characterization (SIC) pipeline providing robust cluster matching and data visualization tools for high-dimensional flow/mass cytometry (and other) data. This pipeline automatically (and intuitively) generates two-dimensional representations of high-dimensional datasets that are safe from the curse of dimensionality. This new approach allows more robust and reproducible data analysis,+ facilitating the development of new gold standard practices across laboratories and institutions.

EuroFlow-Based Flowcytometric Diagnostic Screening and Classification of Primary Immunodeficiencies of the Lymphoid System.

Guidelines for screening for primary immunodeficiencies (PID) are well-defined and several consensus diagnostic strategies have been proposed. These consensus proposals have only partially been implemented due to lack of standardization in laboratory procedures, particularly in flow cytometry. The main objectives of the EuroFlow Consortium were to innovate and thoroughly standardize the flowcytometric techniques and strategies for reliable and reproducible diagnosis and classification of PID of the lymphoid system. The proposed EuroFlow antibody panels comprise one orientation tube and seven classification tubes and corresponding databases of normal and PID samples. The 8-color 12-antibody PID Orientation tube (PIDOT) aims at identification and enumeration of the main lymphocyte and leukocyte subsets; this includes naïve pre-germinal center (GC) and antigen-experienced post-GC memory B-cells and plasmablasts. The seven additional 8(-12)-color tubes can be used according to the EuroFlow PID algorithm in parallel or subsequently to the PIDOT for more detailed analysis of B-cell and T-cell subsets to further classify PID of the lymphoid system. The Pre-GC, Post-GC, and immunoglobulin heavy chain (IgH)-isotype B-cell tubes aim at identification and enumeration of B-cell subsets for evaluation of B-cell maturation blocks and specific defects in IgH-subclass production. The severe combined immunodeficiency (SCID) tube and T-cell memory/effector subset tube aim at identification and enumeration of T-cell subsets for assessment of T-cell defects, such as SCID. In case of suspicion of antibody deficiency, PIDOT is preferably directly combined with the IgH isotype tube(s) and in case of SCID suspicion (e.g., in newborn screening programs) the PIDOT is preferably directly combined with the SCID T-cell tube. The proposed ≥8-color antibody panels and corresponding reference databases combined with the EuroFlow PID algorithm are designed to provide fast, sensitive and cost-effective flowcytometric diagnosis of PID of the lymphoid system, easily applicable in multicenter diagnostic settings world-wide.

PS1234 APPLICATION OF FLOW SOM TO CLASSICAL FLOW CYTOMETRY: UNSUPERVISED ANALYSIS OF HEMATOPOIESIS AND HEMATOLOGICAL MALIGNANCIES

Background:Multiparameter flow cytometry (MFC) is an essential tool for the diagnosis and follow‐up of hematological malignancies. Numerous panels have been devised that best allow to achieve these goals. However, analysis of MFC graphs still heavily relies on the expertise of biologists and on specifically well adapted cascade‐gating strategies. Because of the development of mass cytometry, which generates data in a multidimensional environment of up to 50 dimensions, new strategies have been developed, that could be applied to classical MFC.Aims:In this study, the capacities of a rapid new tool for multidimensional analysis, FlowSOM (self‐organizing maps) have been combined to a classical MFC analysis tool, Kaluza®, for the identification of normal and diseased human bone marrow (BM) cell subpopulations.Methods:Normal BM samples collected during bone surgery were analyzed in MFC with four antibody panels adapted to the exploration of acute myeloid leukemia (AML) and acute lymphoblastic leukemia (ALL) respectively. The listmode files obtained were merged and submitted to FlowSOM analysis, using R packages, which yielded trees of 100 nodes defining normal BM cell subsets. FlowSOM parameters were then transferred to Kaluza® for a bidimensional representation of the trees within which selected navigation could be performed with Kaluza® tools. In a second part of the study, the same stainings were performed on myelodysplastic or leukemic (AML or ALL) samples (diagnosis or follow‐up), which were submitted to FlowSOM together with the reference merged BM.Results:Node‐by‐node selection allowed for a precise definition of each of the unsupervised subsets generated by the software. This disclosed a selective segregation of delicate subsets such as non‐classical monocytes, hematogones or stem cells. Direct simultaneous comparison of diseased samples and normal BM allowed for a precise definition of abnormal cells (blasts) at diagnosis, and their immediate recognition as minimal residual disease in follow‐up samples.Summary/Conclusion:Flow SOM appears as a new and efficient tool for a rapid identification of BM cell subsets, highlighting differences from normal BM in dysplastic of leukemic samples. The node‐by‐node approach facilitates the definition of abnormal cells and their quantification, supervision intervening only after the generation of subsets delineated by artificial intelligence.

PS1235 RED BLOOD CELL GENERATION FROM HUMAN INDUCED PLURIPOTENT STEM CELLS

Background: Multiparameter flow cytometry (MFC) is an essential tool for the diagnosis and follow-up of hematological malignancies. Numerous panels have been devised that best allow to achieve these goals. However, analysis of MFC graphs still heavily relies on the expertise of biologists and on specifically well adapted cascade-gating strategies. Because of the development of mass cytometry, which generates data in a multidimensional environment of up to 50 dimensions, new strategies have been developed, that could be applied to classical MFC. Aims: In this study, the capacities of a rapid new tool for multidimensional analysis, FlowSOM (self-organizing maps) have been combined to a classical MFC analysis tool, Kaluza®, for the identification of normal and diseased human bone marrow (BM) cell subpopulations. Methods: Normal BM samples collected during bone surgery were analyzed in MFC with four antibody panels adapted to the exploration of acute myeloid leukemia (AML) and acute lymphoblastic leukemia (ALL) respectively. The listmode files obtained were merged and submitted to FlowSOM analysis, using R packages, which yielded trees of 100 nodes defining normal BM cell subsets. FlowSOM parameters were then transferred to Kaluza® for a bidimensional representation of the trees within which selected navigation could be performed with Kaluza® tools. In a second part of the study, the same stainings were performed on myelodysplastic or leukemic (AML or ALL) samples (diagnosis or follow-up), which were submitted to FlowSOM together with the reference merged BM. Results: Node-by-node selection allowed for a precise definition of each of the unsupervised subsets generated by the software. This disclosed a selective segregation of delicate subsets such as non-classical monocytes, hematogones or stem cells. Direct simultaneous comparison of diseased samples and normal BM allowed for a precise definition of abnormal cells (blasts) at diagnosis, and their immediate recognition as minimal residual disease in follow-up samples. Summary/Conclusion: Flow SOM appears as a new and efficient tool for a rapid identification of BM cell subsets, highlighting differences from normal BM in dysplastic of leukemic samples. The node-by-node approach facilitates the definition of abnormal cells and their quantification, supervision intervening only after the generation of subsets delineated by artificial intelligence.