Deconvolution Tool Research Articles

Benign breast tumors may arise on different immunological backgrounds.

Benign breast tumors are a nonthreatening condition defined as abnormal cell growth within the breast without the ability to invade nearby tissue. However, benign lesions hold valuable biological information that can lead us toward better understanding of tumor biology. In this study, we have used two pathway analysis algorithms, Pathifier and gene set variation analysis (GSVA), to identify biological differences between normal breast tissue, benign tumors and malignant tumors in our clinical dataset. Our results revealed that one-third of all pathways that were significantly different between benign and malignant tumors were immune-related pathways, and 227 of them were validated by both methods and in the METABRIC dataset. Furthermore, five of these pathways (all including genes involved in cytokine and interferon signaling) were related to overall survival in cancer patients in both datasets. The cellular moieties that contribute to immune differences in malignant and benign tumors were analyzed using the deconvolution tool, CIBERSORT. The results showed that levels of some immune cells were specifically higher in benign than in malignant tumors, and this was especially the case for resting dendritic cells and follicular T-helper cells. Understanding the distinct immune profiles of benign and malignant breast tumors may aid in developing noninvasive diagnostic methods to differentiate between them in the future.

ProteoMixture: A cell type deconvolution tool for bulk tissue proteomic data

Numerous multi-omic investigations of cancer tissue have documented varying and poor pairwise transcript:protein quantitative correlations, and most deconvolution tools aiming to predict cell type proportions (cell admixture) have been developed and credentialed using transcript-level data alone. To estimate cell admixture using protein abundance data, we analyzed proteome and transcriptome data generated from contrived admixtures of tumor, stroma, and immune cell models or those selectively harvested from the tissue microenvironment by laser microdissection from high grade serous ovarian cancer (HGSOC) tumors. Co-quantified transcripts and proteins performed similarly to estimate stroma and immune cell admixture (r ≥ 0.63) in two commonly used deconvolution algorithms, ESTIMATE or ConsensusTME. We further developed and optimized protein-based signatures estimating cell admixture proportions and benchmarked these using bulk tumor proteomic data from over 150 patients with HGSOC. The optimized protein signatures supporting cell type proportion estimates from bulk tissue proteomic data are available at https://lmdomics.org/ProteoMixture/.

CellsFromSpace: a fast, accurate, and reference-free tool to deconvolve and annotate spatially distributed omics data.

Spatial transcriptomics enables the analysis of cell crosstalk in healthy and diseased organs by capturing the transcriptomic profiles of millions of cells within their spatial contexts. However, spatial transcriptomics approaches also raise new computational challenges for the multidimensional data analysis associated with spatial coordinates. In this context, we introduce a novel analytical framework called CellsFromSpace based on independent component analysis (ICA), which allows users to analyze various commercially available technologies without relying on a single-cell reference dataset. The ICA approach deployed in CellsFromSpace decomposes spatial transcriptomics data into interpretable components associated with distinct cell types or activities. ICA also enables noise or artifact reduction and subset analysis of cell types of interest through component selection. We demonstrate the flexibility and performance of CellsFromSpace using real-world samples to demonstrate ICA's ability to successfully identify spatially distributed cells as well as rare diffuse cells, and quantitatively deconvolute datasets from the Visium, Slide-seq, MERSCOPE, and CosMX technologies. Comparative analysis with a current alternative reference-free deconvolution tool also highlights CellsFromSpace's speed, scalability and accuracy in processing complex, even multisample datasets. CellsFromSpace also offers a user-friendly graphical interface enabling non-bioinformaticians to annotate and interpret components based on spatial distribution and contributor genes, and perform full downstream analysis. CellsFromSpace (CFS) is distributed as an R package available from github at https://github.com/gustaveroussy/CFS along with tutorials, examples, and detailed documentation.

Simvastatin Significantly Reduced Alcohol-Induced Cardiac Damage in Adolescent Mice

Alcohol abuse by adolescents is becoming a serious health concern as they often progress to becoming alcoholics later in life which may lead to heart problems. Chronic alcohol use alters the cardiac function and structure, such as haemodynamic changes, weakening and loss of cardiomyocytes, myocardial fibrosis, and inflammation. Simvastatin is a commonly used drug for the treatment and management of various cardiovascular problems but information on its protective effects against alcohol-induced cardiomyocyte hypertrophy, fibrosis, and inflammation is lacking in the literature. Four-week-old male (n = 5) and female (n = 5) C57BL/6 J mice were assigned to each experimental group: (I) NT—no administration of alcohol or Simvastatin; (II) ALC—2.5 g/Kg/day of 20% alcohol via intraperitoneal injection (i.p.); (III) SIM—5 mg/Kg/day of Simvastatin via oral gavage; (iv) ALC + SIM5—5 mg/Kg/day of Simvastatin via oral gavage followed by 2.5 g/Kg/day of 20% alcohol via i.p.; and (v) ALC + SIM15—15 mg/Kg/day Simvastatin via oral gavage followed by 2.5 g/Kg/day of 20% alcohol via i.p. After the 28-day treatment period, the heart was removed and processed for H&E, Masson’s trichrome, or TNF-α immunolabelling. The area and diameter of cardiomyocytes were measured on the H&E-stained sections. The distribution of collagen or TNF-α expression was quantified using the deconvolution tool of ImageJ software. The results confirmed alcohol-induced toxicity on the cardiomyocytes and Simvastatin reduced alcohol-induced cardiomyocyte hypertrophy, fibrosis, and inflammation in both sexes. This study demonstrated that Simvastatin, an FDA approved and easily accessible drug, may be beneficial in lowering the prevalence of alcohol-induced cardiovascular diseases (especially in adolescents) which will have a huge financial implication on health systems worldwide.

Phenotype predictions of two-person mixture using single cell analysis

Over a decade after the publication of the first forensic DNA phenotyping (FDP) studies, DNA-based appearance predictions are now becoming a reality in routine crime scene investigations. The significant number of publications dedicated to the subject of FDP clearly demonstrates a sustained interest and a strong need for further method development. However, the implementation of FDP in routine work still encounters obstacles, and one of these challenges is making phenotype predictions from DNA mixtures. In this study, we examined single-cell sequencing as a potential tool to enable reliable phenotyping of contributors within mixtures. Two mock mixtures, each containing two contributors with similar and different physical appearances, were analyzed using two different workflows. In the first workflow, the mixtures were sequenced using the Ion AmpliSeq™ PhenoTrivium Panel, which includes 41 HIrisPlex-S (HPS) markers. Subsequently, the genotypes were analyzed using the HPS Deconvolution Tool to predict the phenotypes of both contributors. The second workflow involved the introduction of single-cell separation and collection using the DEPArray™ PLUS System. Two different PhenoTrivium amplification protocols were tested, and the phenotype predictions from single cells were compared with the results obtained using the HPS Tool. Our results suggest that the approach presented here allows for the obtainment of nearly complete HIrisPlex-S profiles with accurate genotypes and reliable phenotype predictions from single cells. This method proves successful in deconvoluting mixtures submitted to forensic DNA phenotyping.

Abstract P1021: Estimating Cardiac Cell Composition In Gtex With Single-nucleus Reference Profiles.

Heart failure (HF) is a heterogeneous condition defined by an inability of the heart to meet the circulatory needs of the body. HF is the result of a maladaptive response to chronic stressors during which the heart undergoes structural and cellular remodeling. For example, cardiomyocytes undergo apoptosis after chronic overstimulation while fibroblasts proliferate to fill gaps left by the loss of larger cells. Although changes to the cellular makeup of the heart are known to mediate HF pathology, estimates of basal cardiac cellular composition vary widely. Moreover, despite our understanding of HF as a complex condition that is influenced by environmental and genetic factors, most studies on heart composition only examine small groups of individuals or inbred models. Recent efforts have used single-nucleus RNAseq (snRNAseq) to tally nuclear proportions as a proxy for cellular proportions. However, cellular dissociation biases for certain cell types and skews the resulting estimates. Furthermore, snRNAseq studies are expensive, making them infeasible and underpowered for population-scale analyses. To avoid these issues, we developed a method to estimate cellular composition of bulk RNAseq using cell-type-specific references defined in snRNAseq. While most deconvolution tools are designed for single-cell RNAseq inputs, our method considers the technical and biological differences of snRNAseq. It also allows for the inclusion of additional reference datasets, utilizes simulated controls, and leverages an existing deconvolution tool, MuSiC. We apply our method to the Genome-Tissue-Expression project (GTEx), a multi-tissue database with bulk RNAseq from hundreds of subjects and snRNAseq from a small subset of those subjects. We estimated left ventricular cellular composition in 386 individuals, revealing considerable heterogeneity. We found significant correlations between subject phenotypes and discrete cell populations. This analysis demonstrates our ability to leverage existing, public datasets to accurately map the understudied variability of cardiac cell type variation across human populations. We now look to incorporate genomic datasets to find genetic markers that link composition estimates to population-level variation.

Spatial relationships and immune subsets in the tumor immune microenvironment of head and neck cancers.

6061 Background: The tumor microenvironment provides important insights into cancer behavior and its response to treatment. In particular, the cellular composition and spatial arrangement of tumor immune infiltrates are associated with patient prognosis and survival. We optimized and employed a novel method using readily available hematoxylin and eosin (H&E) slide images, termed the G-cross score, which reflects immune cell infiltration levels and their proximity to tumor cells. The G-cross score is calculated by estimating the probability distribution of immune cells, of any type, within a certain distance from a tumor cell, with higher scores indicating greater immune cell infiltration. However, because G-cross scores are calculated from H&E slides, the method is unable to distinguish among immune cell types. Using head and neck squamous cell carcinoma (HNSCC) patients having both G-cross scores and bulk RNA-seq data, we determined which subset of tumor-infiltrating immune cell types best correlate with G-cross scores, and its predictive and prognostic value. Methods: G-cross scores were calculated for 425 HNSCC samples from University of Michigan, of which 64 had paired bulk RNA-seq data. We applied the cell-type deconvolution tool, CIBERSORTx, to identify immune cell type proportions based on RNA-seq data. G-cross scores were correlated against cell-type deconvolution results, immune cell activation expression signatures, and validated gene expression-based radiosensitivity index scores. Pearson's correlation coefficient and p-value using HPV status as a covariate in a linear model was calculated for each comparison. For survival and recurrence analysis, Cox proportional hazards multivariate regression was performed using the entire 425 sample cohort. Results: G-cross scores were significantly correlated with total T cells (R = 0.52, p = 4.63✕10-5), B cells (R = 0.33, p = 2.08✕10-2), and dendritic cells (R = 0.46, p = 7.71✕10-4), but not macrophages or mast cells. G-cross scores also had a positive correlation with T cell activation (R = 0.46, p = 2.17✕10-4) and B cell activation (R = 0.53, p = 1.24✕10-4) and were correlated with radiation sensitivity (R = -0.44, p = 6.67✕10-4). Survival analysis showed higher G-cross scores were associated with a longer time to recurrence and improved overall survival (p = 3.0✕10-4). Conclusions: G-cross scores are mainly driven by T cell and dendritic cell levels among tumor-infiltrating lymphocytes and are correlated with B and T cell activation. Importantly, G-cross scores also correlate with gene expression profiles associated with response to radiotherapy as well as time to recurrence and overall survival. G-cross scores may be an important biomarker of radiation response and indicative of protective immune subsets for HNSCC patients.

Abstract 5401: Computational cancer cell gene expression deconvolution from tumor bulk RNA-seq via the machine learning algorithm Helenus

Abstract Background: Biomarker gene expression is becoming more commonly utilized for clinical decision-making in oncology clinical practice. However, complex tumor tissue comprises a population of cancer cells (CC) and the tumor microenvironment (TME), causing expression signals belonging to the CC and TME calculated from bulk RNA-seq of the tumor tissue to be indistinguishable. To circumvent this, Helenus, a gene expression deconvolution tool, was developed to estimate TME-specific gene expression, consequently, providing precise CC-specific gene expression. Methods: Helenus performs the “subtraction” of TME gene expression from the total expression calculated from bulk RNA-seq of the tumor tissue. To accurately reconstruct the CC expression profile, LightGBM gene models were trained on artificial transcriptomes created from > 1,000 different solid tumor cancer cell lines and > 3,000 samples of various TME cellular proportions. The LighGBM gene models included genes expressed predominantly in the TME (e.g., CD3E), both the TME and the CC (e.g., BCL6), or in the CC (e.g., HER2). The input features included: 1) RNA percentages of TME cell types predicted by the cell deconvolution tool Kassandra (Zaitsev et al., 2022); 2) evaluation of TME target gene expression via the estimation of its weighted average expression profile in TME cell populations; and 3) a set of TME- and CC-specific genes. The resulting predictions were adjusted based on the CC cell fraction. To evaluate Helenus’ performance, CC and TME RNA were mixed at different ratios using various cancer cell lines and peripheral blood-derived TME cell populations and suspensions of tumor cells prepared from cancer tissue across multiple tumor purity dilutions. Results: Helenus deconvolution resulted in an increased concordance correlation value from 0.73 to 0.98 between the real gene expression profile of pure CC and the reconstructed CC expression from bulk RNA-seq. Helenus showed high concordance between the gene expression profile of sorted cancer cell lines and the deconvolved gene expression across a wide range of CC RNA concentrations (20-90%) mixed with imitated TME RNA at varying concentrations. Helenus demonstrated high performance calculating gene expression of multiple clinically relevant biomarkers in the TME:cancer cell line mixes: CD274 (PD-L1) (mean absolute error [MAE] ~3.5-fold reduction); HLA-A (~2.8-fold MAE reduction); MKI67 (Ki-67, ~2.2-fold MAE reduction), ERBB2 (HER2, ~1.7-fold MAE reduction). Helenus deconvolved CC expression and found significant correlations with CC gene amplifications and deletions (e.g., BCL-2, VNN3) independent of tumor purity (p < 0.003). Conclusion: Helenus, the CC gene expression deconvolution tool, was developed with high accuracy to contribute to tumor diagnosis, disease monitoring, treatment decisions, and clinically relevant biomarker identification. Citation Format: Valentina Beliaeva, Ekaterina Ivleva, Boris Shpak, Daniil Litvinov, Anastasia Zotova, Krystle Nomie, Daniiar Dyikanov, Alexander Kuznetsov, Maria Savchenko, Aleksandr Zaitsev, Nathan Fowler, Alexander Bagaev. Computational cancer cell gene expression deconvolution from tumor bulk RNA-seq via the machine learning algorithm Helenus. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5401.

Abstract 2360: Digital spatial profiling of RNA and protein in the tumor microenvironment of invasive lobular breast cancer

Abstract Invasive lobular breast cancer (ILC) is an understudied breast cancer subtype with late recurrence, metastasis to serosal surfaces including the peritoneum, and dismal long-term outcome. The interaction between a tumor and its microenvironment leads to phenotypic changes in stromal cells and the ECM that promote proliferation and invasion of malignant cells. ILC is histologically distinct from invasive ductal carcinoma, characterized by discohesive E-cadherin-negative tumor cells that grow in single file. We expect the TME to be unique in ILC. We hypothesized that differing levels of nuclear receptor expression in tumor cells would impact stromal cell composition, presumably through paracrine signaling. There is a range of glucocorticoid receptor (GR) expression in ILC. We sought to determine how crosstalk between GR+ or GR- ILC cells and their respective TME differentially impact stromal cell gene and protein expression, as well as the immune cell milieu. To profile the tumor and stroma separately, we performed nanoString GeoMx digital spatial profiling (DSP) of RNA and protein expression in GR-positive (+) and GR-negative (-) primary ILC. We performed RNA DSP for 12 primary ILCs: 5 strongly positive for GR, 5 GR- and 2 tumors with mixed GR expression. We also completed DSP of proteins on adjacent tissue sections for a total of 87 proteins. To profile tumor and TME independently we segmented regions into PanCK+ tumor and PanCK- stroma. RNA expression analyses revealed striking differences between GR+ and GR- ILC in both tumor cells and TME. Pathway analyses showed enrichment of genes encoding eukaryotic translation initiation and control factors and ribosomal proteins in GR+ tumor cells, as well as downregulation of immune response and antigen presentation pathways. In the TME of the GR- tumors, we saw significantly higher expression of collagen biosynthesis and extracellular matrix genes. Using a spatial deconvolution tool, we observed gene expression indicating markedly higher abundance of macrophages and other myeloid subsets, endothelial cells and fibroblasts in GR+ ILC stroma. There was also evidence for more regulatory T-cells, suggestive of an immune suppressive microenvironment in GR+ ILC. Protein DSP revealed upregulation of proteins involved in DNA repair and cell survival, and reduction of some immune checkpoint proteins in the GR+ tumor cells. In the TME of GR+ ILC, we observed more Ki-67 and ER-alpha as well as proteins expressed on T-regs, as well as downregulation of cytotoxic proteins such as granzyme A. We conclude that in-depth examination of molecular profiles in both the tumor cells and the TME of ILCs will give insight into how tumor GR expression/activation influences the crosstalk between ILC cells and their microenvironment. As expected, we observed heterogeneity between patients, but DSP will allow in-depth characterization of individual tumors at the molecular level. Citation Format: Lynda B. Bennett, Sunati Sahoo, Cheryl M. Lewis, Indu Raman, Candace Frerich, Guanchun Chen, Min Xu, Suzanne D. Conzen. Digital spatial profiling of RNA and protein in the tumor microenvironment of invasive lobular breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 2360.

B cell-dependent subtypes and treatment-based immune correlates to survival in stage 3 and 4 lung adenocarcinomas.

Lung cancer is the leading cause of cancer-related deaths worldwide. Surgery and chemoradiation are the standard of care in early stages of non-small cell lung cancer (NSCLC), while immunotherapy is the standard of care in late-stage NSCLC. The immune composition of the tumor microenvironment (TME) is recognized as an indicator for responsiveness to immunotherapy, although much remains unknown about its role in responsiveness to surgery or chemoradiation. In this pilot study, we characterized the NSCLC TME using mass cytometry (CyTOF) and bulk RNA sequencing (RNA-Seq) with deconvolution of RNA-Seq being performed by Kassandra, a recently published deconvolution tool. Stratification of patients based on the intratumoral abundance of B cells identified that the B-cell rich patient group had increased expression of CXCL13 and greater abundance of PD1+ CD8 T cells. The presence of B cells and PD1+ CD8 T cells correlated positively with the presence of intratumoral tertiary lymphoid structures (TLS). We then assessed the predictive and prognostic utility of these cell types and TLS within publicly available stage 3 and 4 lung adenocarcinoma (LUAD) RNA-Seq datasets. As previously described by others, pre-treatment expression of intratumoral 12-chemokine TLS gene signature is associated with progression free survival (PFS) in patients who receive treatment with immune checkpoint inhibitors (ICI). Notably and unexpectedly pre-treatment percentages of intratumoral B cells are associated with PFS in patients who receive surgery, chemotherapy, or radiation. Further studies to confirm these findings would allow for more effective patient selection for both ICI and non-ICI treatments.

Sonicate Fluid Cellularity Predicted by Transcriptomic Deconvolution Differentiates Infectious from Non-Infectious Arthroplasty Failure.

Although cellularity is traditionally assessed morphologically, deep sequencing approaches being used for microorganism detection may be able to provide information about cellularity. We hypothesized that cellularity predicted using CIBERSORTx (Stanford University), a transcriptomic-based cellular deconvolution tool, would differentiate between infectious and non-infectious arthroplasty failure. CIBERSORTx-derived cellularity profiles of 93 sonicate fluid samples, including 53 from subjects who underwent failed arthroplasties due to periprosthetic joint infection (PJI) (abbreviated for the purpose of this study as PJIF) and 40 from subjects who had undergone non-infectious arthroplasty failure (abbreviated NIAF) that had been subjected to bulk RNA sequencing were evaluated. Samples from PJIF and NIAF subjects were differentially clustered by principal component analysis based on the cellularity profile. Twelve of the 22 individual predicted cellular fractions were differentially expressed in the PJIF cases compared with the NIAF cases, including increased predicted neutrophils (mean and standard error, 9.73% ± 1.06% and 0.81% ± 0.60%), activated mast cells (17.12% ± 1.51% and 4.11% ± 0.44%), and eosinophils (1.96% ± 0.37% and 0.42% ± 0.21%), and decreased predicted M0 macrophages (21.33% ± 1.51% and 39.75% ± 2.45%), M2 macrophages (3.56% ± 0.52% and 8.70% ± 1.08%), and regulatory T cells (1.57% ± 0.23% and 3.20% ± 0.34%). The predicted total granulocyte fraction was elevated in the PJIF cases (32.97% ± 2.13% and 11.76% ± 1.61%), and the samples from the NIAF cases had elevated predicted total macrophage and monocyte (34.71% ± 1.71% and 55.34% ± 2.37%) and total B cell fractions (5.89% ± 0.30% and 8.62% ± 0.86%). Receiver operating characteristic curve analysis identified predicted total granulocytes, neutrophils, and activated mast cells as highly able to differentiate between the PJIF cases and the NIAF cases. Within the PJIF cases, the total granulocyte, total macrophage and monocyte, M0 macrophage, and M2 macrophage fractions were differentially expressed in Staphylococcus aureus compared with Staphylococcus epidermidis -associated samples. Within the NIAF cases, the predicted total B cell, naïve B cell, plasma cell, and M2 macrophage fractions were differentially expressed among different causes of failure. CIBERSORTx can predict the cellularity of sonicate fluid using transcriptomic data, allowing for the evaluation of the underlying immune response during the PJIF and NIAF cases, without a need to phenotypically assess cell composition.

An Interpretable Machine-learning Framework for Modeling High-resolution Spectroscopic Data* * Open-source code at https://github.com/gully/blase.

Comparison of échelle spectra to synthetic models has become a computational statistics challenge, with over 10,000 individual spectral lines affecting a typical cool star échelle spectrum. Telluric artifacts, imperfect line lists, inexact continuum placement, and inflexible models frustrate the scientific promise of these information-rich data sets. Here we debut an interpretable machine-learning framework blasé that addresses these and other challenges. The semiempirical approach can be viewed as “transfer learning”—first pretraining models on noise-free precomputed synthetic spectral models, then learning the corrections to line depths and widths from whole-spectrum fitting to an observed spectrum. The auto-differentiable model employs back-propagation, the fundamental algorithm empowering modern deep learning and neural networks. Here, however, the 40,000+ parameters symbolize physically interpretable line profile properties such as amplitude, width, location, and shape, plus radial velocity and rotational broadening. This hybrid data-/model-driven framework allows joint modeling of stellar and telluric lines simultaneously, a potentially transformative step forward for mitigating the deleterious telluric contamination in the near-infrared. The blasé approach acts as both a deconvolution tool and semiempirical model. The general-purpose scaffolding may be extensible to many scientific applications, including precision radial velocities, Doppler imaging, chemical abundances for Galactic archeology, line veiling, magnetic fields, and remote sensing. Its sparse-matrix architecture and GPU acceleration make blasé fast. The open-source PyTorch-based code blase includes tutorials, Application Programming Interface documentation, and more. We show how the tool fits into the existing Python spectroscopy ecosystem, demonstrate a range of astrophysical applications, and discuss limitations and future extensions.

Diagnostic Performance of a Noninvasive Breath Test for Colorectal Cancer: COBRA1 Study

Diagnostic Performance of a Noninvasive Breath Test for Colorectal Cancer: COBRA1 Study

Abstract 5031: Improved deconvolution of combined bulk and single-cell RNA-sequencing data

Abstract Identifying how cell types and their abundances evolve during tumor progression is critical to understanding the mechanisms and identifying predictors of metastasis. Single-cell RNA sequencing (scRNA-seq) has been especially promising in resolving heterogeneity of expression programs at the single cell level but is not always available, for example for large cohort studies or longitudinal analysis of archived samples. In such cases, cell subpopulations must be inferred by deconvolution, a process that can infer single-cell genomic data from bulk data but has limited ability to resolve fine clonal structure. We extend our previous bulk genomic deconvolution tool, Robust and Accurate Deconvolution (RAD), to establish a new method, scRAD, that can use reference scRNA-seq to interpret sample collections for which only bulk RNA-seq is available for some samples, e.g., clonally resolving archived primary (PRM) tissues using scRNA-seq from metastases (METs). We preprocess scRNA-seq data to accurately represent gene expression profiles, yielding a signature matrix S then extend our RAD method via a regularization term to deconvolve bulk data while maximizing consistency with S. We validate our method on semi-synthetic data derived from human PRM breast cancer cases and bone and ovary METs, showing that scRAD improves inference of single-cell gene expression profiles and their frequencies relative to the prior RAD with random initialization or initialization using the single-cell matrix S (Table 1). We then apply scRAD to a collection of paired PRM and MET tumors to quantify progression changes in common cell types. One-sided Kaplan-Meier analysis shows that tumors inferred to increase the mast cell fraction from PRM to MET exhibit lower overall survival (p<0.05), consistent with the role of mast cells in metastatic growth and propagation. Tumors that show increased macrophage cell fraction from PRM to MET show improved overall survival (p<0.04), consistent with the role of immune infiltration in survival. mean square error (MSE) of gene expression and mixture fraction inference on semi-simulated data Method RAD with random initialization RAD with random initialization RAD with random initialization RAD initialized with S RAD initialized with S RAD initialized with S scRAD scRAD scRAD Sample number 2 4 8 2 4 8 2 4 8 Gene Expression MSE 0.41 0.28 0.37 0.31 0.27 0.37 0.22 0.15 0.13 Mixture fraction MSE 0.58 0.82 0.59 0.39 0.83 0.58 0.19 0.20 0.22 Citation Format: Haoyun Lei, Xiaoyan A. Guo, Yifeng Tao, Kai Ding, Xuecong Fu, Steffi Oesterreich, Adrian V. Lee, Russell Schwartz. Improved deconvolution of combined bulk and single-cell RNA-sequencing data [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 5031.

A novel strategy using optimized MOMED and B-spline based envelope-derivative operator for compound fault detection of the rolling bearing

The bearing regularly suffers from compound faults in real-world working conditions. In comparison to the single-fault feature extraction, the compound fault diagnosis is more difficult to achieve. This paper suggests an alternative signal processing strategy using the Multipoint Optimal Minimum Entropy Deconvolution method (MOMED) and B-spline based envelope-derivative operator (EDO) tools. As an upgraded version of the Minimum Entropy Deconvolution tool, the MOMEDA technique has been extensively available for bearing and gear fault detection. However, this approach results in an open problem related to how one can choose an appropriate filter size. Considering this problem, an optimized MOMED based on Salp Swarm Algorithm is proposed. Besides, a novel energy operator method called B-spline based envelope-derivative operator (B-spline EDO) is proposed to detect the corresponding fault characteristics from the two separated mono-component signals produced by the optimized MOMED. The new B-spline EDO method accomplishes higher fault detection performance in a noisy environment. Finally, the experimental results displayed that the novel compound fault detection approach can effectively identify the compound fault characteristics.

Mass Deconvolution of Top-Down Mass Spectrometry Datasets by FLASHDeconv.

Mass deconvolution, the determination of proteoform precursor and fragment masses, is crucial for top-down proteomics data analysis. Here we describe the detailed procedure to run FLASHDeconv, an ultrafast, high-quality mass deconvolution tool. Both spectrum- and feature-level deconvolution results are obtainable in various output formats by FLASHDeconv. FLASHDeconv is runnable in different environments such as the command line and OpenMS workflows.

Single cell transcriptomics in human osteoarthritis synovium and in silico deconvoluted bulk RNA sequencing

To reveal the heterogeneity of different cell types of osteoarthritis (OA) synovial tissues at a single-cell resolution, and determine by novel methodology whether bulk-RNA-seq data could be deconvoluted to create in silico scRNA-seq data for synovial tissue analyses. OA scRNA-seq data (102,077 synoviocytes) were provided by 17 patients undergoing total knee arthroplasty; 9 tissues with matched scRNA-seq and bulk RNA-seq data were used to evaluate six in silico gene deconvolution tools. Predicted and observed cell types and proportions were compared to identify the best deconvolution tool for synovium. We identified seven distinct cell types in OA synovial tissues. Gene deconvolution identified three (of six) platforms as suitable for extrapolating cellular gene expression from bulk RNA-seq data. Using paired scRNA-seq and bulk RNA-seq data, an "arthritis" specific signature matrix was created and validated to have a significantly better predictive performance for synoviocytes than a default signature matrix. Use of the machine learning tool, Cell-type Identification By Estimating Relative Subsets of RNA Transcripts x (CIBERSORTx), to analyze rheumatoid arthritis (RA) and OA bulk RNA-seq data yielded proportions of T cells and fibroblasts that were similar to the gold standard observations from RA and OA scRNA-seq data, respectively. This novel study revealed heterogeneity of synovial cell types in OA and the feasibility of gene deconvolution for synovial tissue.

Bulk Gene Expression Deconvolution Reveals Infiltration of M2 Macrophages in Retinal Neovascularization.

PurposeThis study interrogated the transcriptional features and immune cellular landscape of the retinae of rats subjected to oxygen-induced retinopathy (OIR).MethodsBulk RNA sequencing was performed with retinal RNA isolated from control and OIR rats. Gene set enrichment analysis (GSEA) was undertaken to identify gene sets associated with immune responses in retinal neovascularization. Bulk gene expression deconvolution analysis by CIBERSORTx was performed to identify immune cell types involved in retinal neovascularization, followed by functional enrichment analysis of differentially expressed genes (DEGs). Protein–protein interaction analysis was performed to predict the hub genes relevant to identified immune cell types. CIBERSORTx was applied to profile immune cell types in the macula of patients with both proliferative diabetic retinopathy (PDR) and diabetic macular edema using a public RNA-seq dataset.ResultsTranscriptome analysis by GSEA revealed that the retina of OIR rats and patients with PDR is characterized by increased immunoregulatory interactions and complement cascade. Deconvolution analysis demonstrated that M2 macrophages infiltrate the retinae of OIR rats and patients with PDR. Functional enrichment analysis of DEGs in OIR rats showed that the dysregulated genes are related to leukocyte-mediated immunity and myeloid leukocyte activation. Downstream protein–protein interaction analysis revealed that several potential hub genes, including Ccl2, Itgam, and Tlr2, contribute to M2 macrophage infiltration in the ischemic retina.ConclusionsThis study highlights application of the gene expression deconvolution tool to identify immune cell types in inflammatory ocular diseases with transcriptomes, providing a new approach to assess changes in immune cell types in diseased ocular tissues.

MMDIT: A tool for the deconvolution and interpretation of mitochondrial DNA mixtures

Short tandem repeats of the nuclear genome have been the preferred markers for analyzing forensic DNA mixtures. However, when nuclear DNA in a sample is degraded or limited, mitochondrial DNA (mtDNA) markers provide a powerful alternative. Though historically considered challenging, the interpretation and analysis of mtDNA mixtures have recently seen renewed interest with the advent of massively parallel sequencing. However, there are only a few software tools available for mtDNA mixture interpretation. To address this gap, the Mitochondrial Mixture Deconvolution and Interpretation Tool (MMDIT) was developed. MMDIT is an interactive application complete with a graphical user interface that allows users to deconvolve mtDNA (whole or partial genomes) mixtures into constituent donor haplotypes and estimate random match probabilities on these resultant haplotypes. In cases where deconvolution might not be feasible, the software allows mixture analysis directly within a binary framework (i.e. qualitatively, only using data on allele presence/absence). This paper explains the functionality of MMDIT, using an example of an in vitro two-person mtDNA mixture with a ratio of 1:4. The uniqueness of MMDIT lies in its ability to resolve mixtures into complete donor haplotypes using a statistical phasing framework before mixture analysis and evaluating statistical weights employing a novel graph algorithm approach. MMDIT is the first available open-source software that can automate mtDNA mixture deconvolution and analysis. The MMDIT web application can be accessed online at https://www.unthsc.edu/mmdit/. The source code is available at https://github.com/SammedMandape/MMDIT_UI and archived on zenodo (https://doi.org/10.5281/zenodo.4770184).

Abstract 3179: Deconvolving bulk gene expression to chart the immune activity and crosstalk in the tumor microenvironment

Abstract The tumor microenvironment is a complex mixture of cell-types that interact with each other to affect tumor growth and clinical outcomes. To accelerate the discovery of such interactions, we develop CODEFACS (Confident Deconvolution For All Cell Subsets), a deconvolution tool that infers cell-type-specific gene expression in each sample from bulk expression measurements. We additionally developed LIRICS (Ligand Receptor Interactions between Cell Subsets), a pipeline that analyzes the deconvolved gene expression to further identify clinically relevant ligand-receptor interactions between cell-types. Using 15 benchmark datasets, we first demonstrate that CODEFACS substantially improves our ability to reconstruct cell-type-specific transcriptomes from individual bulk samples compared to the state-of-the-art method, CIBERSORTx. We next analyze the TCGA using CODEFACS and LIRICS to uncover tumor-immunological cell-type-specific interactions that are differentially activate in mismatch repair deficient tumors. This analysis shows that T-cell co-stimulating interactions are associated with higher survival rates in these tumors independent of their mutation burden levels. Finally, using machine learning, we identify key cell-type-specific ligand-receptor interactions that predict patient response to immune checkpoint blockade therapy in melanoma. Taken together, we present and validate a computational framework for analyzing large bulk gene expression datasets at a cell-type specific resolution that has many potential future applications, complementing single cell transcriptomics in a cost-effective manner. Citation Format: Kun Wang, Sushant Patkar, JooSang Lee, E. Michael Gertz, Welles Robinson, Fiorella Schischlik, Alejandro Schäffer, Eytan Ruppin. Deconvolving bulk gene expression to chart the immune activity and crosstalk in the tumor microenvironment [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 3179.