Differential expression analysis with inmoose, the integrated multi-omic open-source environment in Python

BackgroundDifferential expression and feature selection analyses are essential steps for the development of accurate diagnostic/prognostic classifiers of complicated human diseases using transcriptomics data. These steps are particularly challenging due to the curse of dimensionality and the presence of technical and biological noise. A promising strategy for overcoming these challenges is the incorporation of pre-existing transcriptomics data in the identification of differentially expressed (DE) genes. This approach has the potential to improve the quality of selected genes, increase classification performance, and enhance biological interpretability. While a number of methods have been developed that use pre-existing data for differential expression analysis, existing methods do not leverage the identities of experimental conditions to create a robust metric for identifying DE genes.ResultsIn this study, we propose a novel differential expression and feature selection method—GEOlimma—which combines pre-existing microarray data from the Gene Expression Omnibus (GEO) with the widely-applied Limma method for differential expression analysis. We first quantify differential gene expression across 2481 pairwise comparisons from 602 curated GEO Datasets, and we convert differential expression frequencies to DE prior probabilities. Genes with high DE prior probabilities show enrichment in cell growth and death, signal transduction, and cancer-related biological pathways, while genes with low prior probabilities were enriched in sensory system pathways. We then applied GEOlimma to four differential expression comparisons within two human disease datasets and performed differential expression, feature selection, and supervised classification analyses. Our results suggest that use of GEOlimma provides greater experimental power to detect DE genes compared to Limma, due to its increased effective sample size. Furthermore, in a supervised classification analysis using GEOlimma as a feature selection method, we observed similar or better classification performance than Limma given small, noisy subsets of an asthma dataset.ConclusionsOur results demonstrate that GEOlimma is a more effective method for differential gene expression and feature selection analyses compared to the standard Limma method. Due to its focus on gene-level differential expression, GEOlimma also has the potential to be applied to other high-throughput biological datasets.

The confounding effects of high genetic diversity on the determination and interpretation of differential gene expression analysis in the parasitic nematode Haemonchus contortus

BackgroundEpigenome-wide association studies (EWAS) and differential gene expression analyses are generally performed on tissue samples, which consist of multiple cell types. Cell-type-specific effects of a trait, such as disease, on the omics expression are of interest but difficult or costly to measure experimentally. By measuring omics data for the bulk tissue, cell type composition of a sample can be inferred statistically. Subsequently, cell-type-specific effects are estimated by linear regression that includes terms representing the interaction between the cell type proportions and the trait. This approach involves two issues, scaling and multicollinearity.ResultsFirst, although cell composition is analyzed in linear scale, differential methylation/expression is analyzed suitably in the logit/log scale. To simultaneously analyze two scales, we applied nonlinear regression. Second, we show that the interaction terms are highly collinear, which is obstructive to ordinary regression. To cope with the multicollinearity, we applied ridge regularization. In simulated data, nonlinear ridge regression attained well-balanced sensitivity, specificity and precision. Marginal model attained the lowest precision and highest sensitivity and was the only algorithm to detect weak signal in real data.ConclusionNonlinear ridge regression performed cell-type-specific association test on bulk omics data with well-balanced performance. The omicwas package for R implements nonlinear ridge regression for cell-type-specific EWAS, differential gene expression and QTL analyses. The software is freely available from https://github.com/fumi-github/omicwas

The insulin sensitizer drug, rosiglitazone, has been shown to have a protective effect on pancreatic islet cell structure and function in animal models of type 2 diabetes. The identification of new molecular targets associated both with islet cell dysfunction and protection is a crucial research goal. In the present study, a proteomics approach has been used to identify such targets. Obese C57Bl/6J lep/lep mice and lean littermates were given the insulin sensitizer drug BRL49653, rosiglitazone. It normalized the impaired glucose tolerance in lep/lep mice but had no significant effect on glucose tolerance in the lean mice. Pancreatic islet polypeptides were arrayed by a two-dimensional gel electrophoresis system that separated more than 2500 individual spots. Three overexpressed and six underexpressed proteins were significant (p < 0.05) between lep/lep and lean mice, and four were modulated significantly (p < 0.05) by the rosiglitazone treatment of the obese mice. The identity of these differentially expressed proteins was made using mass spectrometric analysis and provided evidence that differential expression of actin-binding proteins may be an important aspect of defective islet function. Rosiglitazone increased carboxypeptidase B expression in both lep/lep and normal mice suggesting that this might be an independent effect of rosiglitazone that contributes to improved insulin processing.

e17601 Background: The oral antioxidant pterostilbene (PT) (3,5-dimethoxy-40-hydroxystilbene) has demonstrated synergistic antiproliferative effect with megestrol acetate (MA) in endometrial cancer (EC). We conducted a preoperative (window of opportunity) study enrolling endometrial cancer patients scheduled for hysterectomy, who were randomized to preoperative PT+MA vs MA alone. Methods: This prospective, randomized phase II clinical trial (NCT03671811) enrolled patients who were surgical candidates for hysterectomy. Patients were randomized 1:1 to each arm and were treated for 3 weeks with either PT+MA (arm 1) or MA alone (arm 2). Exploratory objectives included histologic response by GOG 211 criteria, and transcriptional changes using whole transcriptome analysis. For the RNA-seq analysis, differential expression analysis used the “limma” R limma package. The differentially expressed genes were selected based on the cutoff values of p-value &lt;= 0.05 and fold change &gt;= 1.5. The gene set enrichment analysis was performed using GSEA software v. 4.3.2 to identify the affected Hallmark and KEGG pathways, using the pre-ranked gene list ranked by the -log10(p-value) with a sign determined by the fold change direction. Results: This prospective, randomized phase II clinical trial (NCT03671811) enrolled patients who were surgical candidates for hysterectomy. Patients were randomized 1:1 to each arm and were treated for 3 weeks with either PT+MA (arm 1) or MA alone (arm 2). Exploratory objectives included histologic response by GOG 211 criteria, and transcriptional changes using whole transcriptome analysis. For the RNA-seq analysis, differential expression analysis used the “limma” R limma package. The differentially expressed genes were selected based on the cutoff values of p-value &lt;= 0.05 and fold change &gt;= 1.5. The gene set enrichment analysis was performed using GSEA software v. 4.3.2 to identify the affected Hallmark and KEGG pathways, using the pre-ranked gene list ranked by the -log10(p-value) with a sign determined by the fold change direction. Conclusions: In endometrial cancer, the oral antioxidant Pterostilbene increases innate immune system pathway expression, while reducing mTOR pathway expression, potentially serving as a useful adjunct to progestin hormonal therapy in EC. Clinical trial information: NCT03671811 .

Identification of a Relapse Signature in t(8;21) Pediatric AML

Differential gene expression analysis is one of the significant efforts in single cell RNA sequencing (scRNAseq) analysis to discover the specific changes in expression levels of individual cell types. Since scRNAseq exhibits multimodality, large amounts of zero counts, and sparsity, it is different from the traditional bulk RNA sequencing (RNAseq) data. The new challenges of scRNAseq data promote the development of new methods for identifying differentially expressed (DE) genes. In this study, we proposed a new method, SigEMD, that combines a logistic regression model and a nonparametric method based on Earth Mover's Distance, to precisely and efficiently identify DE genes in scRNAseq data. The regression model is used to reduce the impact of large amounts of zero counts, and the nonparametric method is used to improve the sensitivity of detecting DE genes from multimodal scRNAseq data. By additionally employing gene interaction network information to adjust the final states of DE genes, we further reduce the false positives of calling DE genes. We used simulated data and real data to evaluate the detection accuracy of the proposed method and to compare its performance with those of other differential expression analysis methods. Results indicate that the proposed method has an overall powerful performance in terms of precision in detection, sensitivity, and specificity.

Study question Can higher concentrations of di-2-ethylhexyl phthalate (DEHP) in follicular fluid (FF) lead to an altered microRNA expression profile? Summary answer The present study indicates the upregulation of miR-203a-3p, miR-150-5p and miR-28-3p in the follicular fluid of women with higher DEHP concentrations (FDR&amp;lt;0.05). What is known already Women are exposed daily to endocrine-disrupting chemicals (EDCs) such as phthalates which can be found in hundreds of personal care products and plastic items. In studies of rodents exposed to certain phthalates, high doses have been shown to change hormone levels, disrupt folliculogenesis and cause birth defects. Recently, we reported a strong inverse association between DEHP metabolites and ovarian sensitivity index (OSI) in females undergoing ART treatment. Several mechanisms have been proposed to explain the association between DEHP and fertility in rodents, however not much is known about the mechanisms in humans. Study design, size, duration FF was collected from 96 women undergoing IVF treatment after controlled ovarian stimulation during ovarian puncture. 48 participants from Estonia were recruited at Nova Vita Clinic AS in Tallinn, Estonia in 2019 and 48 participants from Sweden were recruited at Carl von Linnekliniken in Uppsala, Sweden in 2016. In both cohorts, patients signed a written informed consent form. Both cohorts were matched for age, BMI, parity and the number of previous IVF treatments. Participants/materials, setting, methods DEHP metabolite concentrations in the FF samples were measured in our earlier study by LC-MS/MS. Cell free RNA was extracted from FF samples with Qiagen miRNeasy Micro kit, sequencing libraries were prepared with QIAseq miRNA library kit and sequenced on Illumina NextSeq 550 instrument. Differential miRNA expression analysis was performed with the DESeq2 package. Pathway enrichment analysis for differentially expressed miRNA targets was conducted with the miEAA tool integrated with WikiPathways. Main results and the role of chance In total, 465 miRNAs were observed at least twice in ≥ 24 samples. Differential expression analysis was conducted between DEHP-high and DEHP-low groups in combined cohorts using the cohort as a covariate. 16 miRNAs were found to be differentially expressed between DEHP groups with FDR&amp;lt;0.1. Three miRNAs: miR-203a-3p, miR-150-5p and miR-28-3p, were differentially expressed with FDR&amp;lt;0.05 and were more abundant in women with high DEHP-levels in both cohorts. Of these, miR-203a-3p was negatively correlated with OSI (Pearson R = -0.27, p = 0.0089). Six miRNAs (miR-10b-5p, miR-112-5p, miR-132-5p, miR-203-3p, miR-205-5p and miR-27-3p) were found to have predicted targets in the androgen receptor signalling pathway. Limitations, reasons for caution The miRNAs found in this study need to be validated in a larger set of patients as a cohort-dependent difference in the results was observed. There is a need for further functional studies to investigate miRNAs as possible mediators of the association between DEHP and ovarian sensitivity index. Wider implications of the findings The current study presents evidence that DEHP exposure alters the expression of FF miRNAs which are associated with androgen receptor signalling. These results provide novel evidence of the potential influence of endocrine-disrupting chemicals on female reproduction. Trial registration number NA

BackgroundNormalization is an important data preparation step in gene expression analyses, designed to remove various systematic noise. Sample variance is greatly reduced after normalization, hence the power of subsequent statistical analyses is likely to increase. On the other hand, variance reduction is made possible by borrowing information across all genes, including differentially expressed genes (DEGs) and outliers, which will inevitably introduce some bias. This bias typically inflates type I error; and can reduce statistical power in certain situations. In this study we propose a new differential expression analysis pipeline, dubbed as super-delta, that consists of a multivariate extension of the global normalization and a modified t-test. A robust procedure is designed to minimize the bias introduced by DEGs in the normalization step. The modified t-test is derived based on asymptotic theory for hypothesis testing that suitably pairs with the proposed robust normalization.ResultsWe first compared super-delta with four commonly used normalization methods: global, median-IQR, quantile, and cyclic loess normalization in simulation studies. Super-delta was shown to have better statistical power with tighter control of type I error rate than its competitors. In many cases, the performance of super-delta is close to that of an oracle test in which datasets without technical noise were used. We then applied all methods to a collection of gene expression datasets on breast cancer patients who received neoadjuvant chemotherapy. While there is a substantial overlap of the DEGs identified by all of them, super-delta were able to identify comparatively more DEGs than its competitors. Downstream gene set enrichment analysis confirmed that all these methods selected largely consistent pathways. Detailed investigations on the relatively small differences showed that pathways identified by super-delta have better connections to breast cancer than other methods.ConclusionsAs a new pipeline, super-delta provides new insights to the area of differential gene expression analysis. Solid theoretical foundation supports its asymptotic unbiasedness and technical noise-free properties. Implementation on real and simulated datasets demonstrates its decent performance compared with state-of-art procedures. It also has the potential of expansion to be incorporated with other data type and/or more general between-group comparison problems.

Development of Poly(A)-ClickSeq as a tool enabling simultaneous genome-wide poly(A)-site identification and differential expression analysis.

BRCAness is a phenotypical trait resembling tumors with germline BRCA1/BRCA2 mutation and homologous recombination deficiency (HRD). Recently, osteosarcoma (OS), which is the most common primary malignant bone tumor, was shown to frequently present molecular features resembling BRCAness. The inability of the HR pathway to keep the genome integrity leads to its instability, resulting in copy number variations and loss of heterozygosity. Tumors characterized by BRCAness might have an increased sensitivity to poly ADP ribose polymerase inhibitors, which leads to synthetic lethality through tumor cell apoptosis. There are several genomic methods to assess BRCAness such as calculating the HRD score based on large genomic rearrangements. We developed a new approach based on DNA-methylation patterns and gene expression signatures. Low-coverage whole-genome (lcWGS) data, DNA-methylation, and RNA-Seq data of 41 patients with relapsed OS were obtained within the scope of the INdividualized Therapy FOr Relapsed Malignancies in Childhood (INFORM) project. Percentage of genome changed (PGC) by large state transitions and telomeric allele imbalance was computed with Nexus CN (BioDiscovery, Inc) and open source Control-FREEC/HRDtools from lcWGS and used for BRCAness scoring. llumina Human Methylation450k and EPIC arrays were used to determine DNA methylation profiles. We selected 18 BRCA-positive and 12 BRCA-negative samples based on the PGC to train the classifier algorithm Random Forest (RF) (R-project) and set aside 11 samples as a test set. Probes were ranked according to minimum variability within groups and maximum between groups. The top 1,000 probes were selected for RF training. The importance of probes was assessed as a mean decrease in accuracy (MDA). Probes with zero MDA were removed from training set. RNA-seq differential expression analysis was performed with DESeq2. Expression signatures were extracted by gene set enrichment analysis (GSEA). Training samples were clustered unambiguously into BRCA-positive and BRCA-negative groups with unsupervised hierarchical clustering of 1,000 probes. All 11 test samples set aside before RF training were accurately classified. Excluding noninformative probes reduced the number of informative probes to 628 and correctly clustered them into two groups. Differential expression analysis of BRCA positive and negative RNA-Seq samples resulted in 1,458 differentially expressed genes. Only upregulated genes in BRCA-positive samples showed highly significant homologous recombination, nucleotide excision, and mismatch repair signatures in GSEA. We established and validated a DNA methylation-based classifier that differentiates BRCA-negative and BRCA-positive OS samples with high accuracy. The differential expression and enrichment analyses identified genome instability signatures. Our method complements existing genomic BRCAness scoring and brings new epigenomic/trancriptomic dimensions into the computational prediction of BRCAness. Citation Format: Maxim Barenboim, Michal Kovac, Baptiste Ameline, David T.W. Jones, Olaf Witt, Stefan Bielack, Stefan Burdach, Daniel Baumhoer, Michaela Nathrath. Identifying BRCAness in osteosarcoma with DNA-methylation profiling and gene expression signatures [abstract]. In: Proceedings of the AACR Special Conference on the Advances in Pediatric Cancer Research; 2019 Sep 17-20; Montreal, QC, Canada. Philadelphia (PA): AACR; Cancer Res 2020;80(14 Suppl):Abstract nr A24.

Radiation Biodosimetry is a crucial method for evaluating the biological consequences of radiation exposure on health outcomes. This technique estimates radiation dose by analyzing the biological effects it causes in the body becomes especially valuable in situations where traditional physical dosimeters, like badges, are unavailable or unreliable. In these scenarios, Radiation Biodosimetry enables accurate dose assessment, which is critical for making informed decisions during emergency response (triage), treatment planning, and long-term health risk evaluation. Current biodosimetry methods struggle to accurately estimate radiation doses, especially for low-dose exposures or complex scenarios involving both neutrons and photons, like those arising from improvised nuclear devices. As a first step towards this goal, we designed and administered tailored radiation doses to human blood samples at two nuclear research facilities, Radiological Research Accelerator Facility at Columbia University (RARAF) and Rhode Island Nuclear Science Center (RINSC), allowing us to analyze the impact of neutrons alongside photons in a controlled setting. The RARAF facility employs a neutron irradiator delivering a broad spectrum of neutron energies, unlike RINSC which utilizes thermal neutrons with a significantly lower energy. This dissertation investigates the impacts of neutron radiation on biological responses using diverse experimental designs. In Chapter 1, to understand the underlying mechanisms of gene expression changes, Differential expression analysis for sequence count data (DESeq2) was utilized to perform the differential gene expression analysis. Three distinct approaches were implemented to achieve a more in-depth understanding. Method 1 prioritized the impact of various radiation doses (Condition) 0.5 Gy, 1 Gy, 2 Gy, and 4 Gy and the radiation source (Batch) at two facilities RARAF and RINSC. Expanding on Method 1, Method 2 approach incorporated the potential influence of sex in addition to the radiation dose and source. Similar to Method 2, Method 3 investigated the potential impact of age as an additional factor alongside condition and source. Our findings demonstrate a significantly greater response in genes differentially expressed by neutron radiation at the RARAF facility compared to exposures at the nuclear reactor, RINSC. This observation highlights unique patterns of gene expression in response to neutron irradiation, with variations influenced by both dose and exposure duration at each facility. The age comparisons in chapters 1 and 2 may be limited by differences in sample size. This could be due to a lack of statistical power to detect age related effects. Future studies with more balanced group sizes are needed to confirm the observed

Microarrays produces enormous amounts of information requiring a series of repeated analyses to condense data. To analyze this data several computational software is used. Objective: To compare the analysis of R and Mathematica package for differential gene expression analysis using microarray dataset. Methods: Microarray Data were collected from an online database GEO (gene expression omnibus). Mathematica and R software was used for comparative analysis. In R software, Robust Multi-Array Average (RMA), was used for data normalization. While Limma package was used for DGE analysis. In Mathematica software, AffyDGED was used for normalization and DGE analysis of dataset. Results: 3,426 non-differentially expressed genes and 14936 genes with differential expression were separated from R. The thresholds for identifying "up" and "down" gene expression were estimated to be 0.98 and -0.19, respectively, using the RMA method to analyze this dataset. AffyDGED from Mathematica detected 1,832 genes as differentially expressed; of them, 1,591 genes overlap with the real and 1,944 differently expressed genes, giving the true positive rate of (1591/1944) =0.818. This indicates that 18% of the genuine list of differentially expressed genes could not be reliably identified by AffyDGED. Conclusions: R programming is one of the most popular and recommendable tools for microarrays to perform different analysis, and along with Bioconductor it makes one of the best analysis algorithms for DGE analysis. On the other hand, AffyDGED brings a contemporary algorithm useful in the real world to the Mathematica user.

RNA-Seq is a widely used technology that allows an efficient genome-wide quantification of gene expressions for, for example, differential expression (DE) analysis. After a brief review of the main issues, methods and tools related to the DE analysis of RNA-Seq data, this article focuses on the impact of both the replicate number and library size in such analyses. While the main drawback of previous relevant studies is the lack of generality, we conducted both an analysis of a two-condition experiment (with eight biological replicates per condition) to compare the results with previous benchmark studies, and a meta-analysis of 17 experiments with up to 18 biological conditions, eight biological replicates and 100 million (M) reads per sample. As a global trend, we concluded that the replicate number has a larger impact than the library size on the power of the DE analysis, except for low-expressed genes, for which both parameters seem to have the same impact. Our study also provides new insights for practitioners aiming to enhance their experimental designs. For instance, by analyzing both the sensitivity and specificity of the DE analysis, we showed that the optimal threshold to control the false discovery rate (FDR) is approximately 2−r, where r is the replicate number. Furthermore, we showed that the false positive rate (FPR) is rather well controlled by all three studied R packages: DESeq, DESeq2, and edgeR. We also analyzed the impact of both the replicate number and library size on gene ontology (GO) enrichment analysis. Interestingly, we concluded that increases in the replicate number and library size tend to enhance the sensitivity and specificity, respectively, of the GO analysis. Finally, we recommend to RNA-Seq practitioners the production of a pilot data set to strictly analyze the power of their experimental design, or the use of a public data set, which should be similar to the data set they will obtain. For individuals working on tomato research, on the basis of the meta-analysis, we recommend at least four biological replicates per condition and 20 M reads per sample to be almost sure of obtaining about 1000 DE genes if they exist.

Introduction: Cancer dissemination involves ongoing dynamic transit of tumor cells between primary and metastatic sites via a hematogenous or “liquid phase”. To identify molecular mechanisms related to the liquid phase of cancer dissemination, we developed a tractable mouse xenograft model wherein GFP-labeled pure primary tumor (PT) cells, circulating tumor cells (CTCs) and lung metastasis (LM) cells are isolated and characterized from the same host. Methods: GFP+ MDA-MB-231 human breast cancer cells were inoculated subcutaneously into NOD/SCID/IL2r-γnull (NSG) mice. When a 1cm diameter primary tumor xenograft formed, 500 μl blood were collected by intracardiac puncture, and CTCs were enriched by wbc/rbc depletion (StemCell Tech). PT and LM samples were fluorescence microdissected from the same mouse hosts and disaggregated into single cell suspensions. Using a micromanipulator pipette, matched PT, CTC, and LM samples (100 pure GFP+ cells each) were isolated, lysed for RNA extraction (Ambion), and analyzed using the Ion AmpliSeq Transcriptome Human Gene Expression workflow for library preparation and pooled sequencing (Ion Proton). Sequence data from biologic triplicates was aligned and curated using the Genetrix bioinformatics suite, and candidate genes were further analyzed using Ingenuity Pathway Analysis (IPA). Results: The sequence panel provides gene-level expression information from over 20,000 genes covering &amp;gt; 95% of the RefSeq gene database. Initially, we focused on genes that were up- or down- regulated in CTCs compared with both PT and LM (i.e. differentially expressed in the liquid phase: 402 genes, 184 up-regulated in CTCs, 218 down-regulated in CTCs, fold change range 1.2x-22x). After further filtering for false discovery, 4 up-regulated and 39 down-regulated genes were selected for IPA analysis, which revealed associations with cellular movement pathways (16 genes) and cancer pathways (24 genes). Of these, the top-ranked 12 differentially expressed genes in CTCs relative to PT and LM (9 down-regulated and 3 up-regulated) were selected for validation by qPCR and further functional characterization in the mouse xenograft model. Conclusions: Here we leveraged new techniques for pure CTC capture, AmpliSeq transcriptome amplification and sequencing to identify and validate new candidate drivers of cancer dissemination. This model holds the potential to uncover key mechanisms engaged by cells that leave the primary and metastatic tumor niche and enter the cancer's liquid phase. Further validation and functional analysis of these candidates in vitro, in vivo and in patient samples may identify new therapeutic targets that specifically disrupt cancer dissemination. Citation Format: Tong Xu, Gareth Morrison, Yucheng Xu, Timothy Triche, Jonathan Buckley, Amir Goldkorn. Differential gene expression analysis by NGS of primary tumor xenograft cells, circulating tumor cells, and metastasis cells purified from the same mouse hosts to identify drivers of cancer dissemination. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 1546.