Dataset Information

Uncovering low-dimensional, miR-based signatures of acute myeloid and lymphoblastic leukemias with a machine-learning-driven network approach.

ABSTRACT: Complex phenotypic differences among different acute leukemias cannot be fully captured by analyzing the expression levels of one single molecule, such as a miR, at a time, but requires systematic analysis of large sets of miRs. While a popular approach for analysis of such datasets is principal component analysis (PCA), this method is not designed to optimally discriminate different phenotypes. Moreover, PCA and other low-dimensional representation methods yield linear or non-linear combinations of all measured miRs. Global human miR expression was measured in AML, B-ALL, and TALL cell lines and patient RNA samples. By systematically applying support vector machines to all measured miRs taken in dyad and triad groups, we built miR networks using cell line data and validated our findings with primary patient samples. All the coordinately transcribed members of the miR-23a cluster (which includes also miR-24 and miR-27a), known to function as tumor suppressors of acute leukemias, appeared in the AML, B-ALL and T-ALL centric networks. Subsequent qRT-PCR analysis showed that the most connected miR in the B-ALL-centric network, miR-708, is highly and specifically expressed in B-ALLs, suggesting that miR-708 might serve as a biomarker for B-ALL. This approach is systematic, quantitative, scalable, and unbiased. Rather than a single signature, our approach yields a network of signatures reflecting the redundant nature of biological signaling pathways. The network representation allows for visual analysis of all signatures by an expert and for future integration of additional information. Furthermore, each signature involves only small sets of miRs, such as dyads and triads, which are well suited for in depth validation through laboratory experiments. In particular, loss-and gain-of-function assays designed to drive changes in leukemia cell survival, proliferation and differentiation will benefit from the identification of multi-miR signatures that characterize leukemia subtypes and their normal counterpart cells of origin.

SUBMITTER: Candia J

PROVIDER: S-EPMC4888874 | biostudies-literature | 2015 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Uncovering low-dimensional, miR-based signatures of acute myeloid and lymphoblastic leukemias with a machine-learning-driven network approach.

Candia Julián J Cherukuri Srujana S Guo Yin Y Doshi Kshama A KA Banavar Jayanth R JR Civin Curt I CI Losert Wolfgang W

Convergent science physical oncology 20151221 2

Complex phenotypic differences among different acute leukemias cannot be fully captured by analyzing the expression levels of one single molecule, such as a miR, at a time, but requires systematic analysis of large sets of miRs. While a popular approach for analysis of such datasets is principal component analysis (PCA), this method is not designed to optimally discriminate different phenotypes. Moreover, PCA and other low-dimensional representation methods yield linear or non-linear combination ...[more]

PMID: 27274862

Similar Datasets

Project description:Automatic boundary detection of 4D ultrasound (4DUS) cardiac data is a promising yet challenging application at the intersection of machine learning and medicine. Using recently developed murine 4DUS cardiac imaging data, we demonstrate here a set of three machine learning models that predict left ventricular wall kinematics along both the endo- and epi-cardial boundaries. Each model is fundamentally built on three key features: (1) the projection of raw US data to a lower dimensional subspace, (2) a smoothing spline basis across time, and (3) a strategic parameterization of the left ventricular boundaries. Model 1 is constructed such that boundary predictions are based on individual short-axis images, regardless of their relative position in the ventricle. Model 2 simultaneously incorporates parallel short-axis image data into their predictions. Model 3 builds on the multi-slice approach of model 2, but assists predictions with a single ground-truth position at end-diastole. To assess the performance of each model, Monte Carlo cross validation was used to assess the performance of each model on unseen data. For predicting the radial distance of the endocardium, models 1, 2, and 3 yielded average R2 values of 0.41, 0.49, and 0.71, respectively. Monte Carlo simulations of the endocardial wall showed significantly closer predictions when using model 2 versus model 1 at a rate of 48.67%, and using model 3 versus model 2 at a rate of 83.50%. These finding suggest that a machine learning approach where multi-slice data are simultaneously used as input and predictions are aided by a single user input yields the most robust performance. Subsequently, we explore the how metrics of cardiac kinematics compare between ground-truth contours and predicted boundaries. We observed negligible deviations from ground-truth when using predicted boundaries alone, except in the case of early diastolic strain rate, providing confidence for the use of such machine learning models for rapid and reliable assessments of murine cardiac function. To our knowledge, this is the first application of machine learning to murine left ventricular 4DUS data. Future work will be needed to strengthen both model performance and applicability to different cardiac disease models.

Project description:BackgroundThe role of Mast cells has not been thoroughly explored in the context of prostate cancer's (PCA) unpredictable prognosis and mixed immunotherapy outcomes. Our research aims to employs a comprehensive computational methodology to evaluate Mast cell marker gene signatures (MCMGS) derived from a global cohort of 1091 PCA patients. This approach is designed to identify a robust biomarker to assist in prognosis and predicting responses to immunotherapy.MethodsThis study initially identified mast cell-associated biomarkers from prostate adenocarcinoma (PRAD) patients across six international cohorts. We employed a variety of machine learning techniques, including Random Forest, Support Vector Machine (SVM), Lasso regression, and the Cox Proportional Hazards Model, to develop an effective MCMGS from candidate genes. Subsequently, an immunological assessment of MCMGS was conducted to provide new insights into the evaluation of immunotherapy responses and prognostic assessments. Additionally, we utilized Gene Set Enrichment Analysis (GSEA) and pathway analysis to explore the biological pathways and mechanisms associated with MCMGS.ResultsMCMGS incorporated 13 marker genes and was successful in segregating patients into distinct high- and low-risk categories. Prognostic efficacy was confirmed by survival analysis incorporating MCMGS scores, alongside clinical parameters such as age, T stage, and Gleason scores. High MCMGS scores were correlated with upregulated pathways in fatty acid metabolism and β-alanine metabolism, while low scores correlated with DNA repair mechanisms, homologous recombination, and cell cycle progression. Patients classified as low-risk displayed increased sensitivity to drugs, indicating the utility of MCMGS in forecasting responses to immune checkpoint inhibitors.ConclusionThe combination of MCMGS with a robust machine learning methodology demonstrates considerable promise in guiding personalized risk stratification and informing therapeutic decisions for patients with PCA.

Project description:BACKGROUND: MicroRNA dysregulation is a common event in leukemia. Polymorphisms in microRNA-binding sites (miRSNPs) in target genes may alter the strength of microRNA interaction with target transcripts thereby affecting protein levels. In this study we aimed at identifying miRSNPs associated with leukemia risk and assessing impact of these miRSNPs on miRNA binding to target transcripts. METHODS: We analyzed with specialized algorithms the 3' untranslated regions of 137 leukemia-associated genes and identified 111 putative miRSNPs, of which 10 were chosen for further investigation. We genotyped patients with acute myeloid leukemia (AML, n?=?87), chronic myeloid leukemia (CML, n?=?140), childhood acute lymphoblastic leukemia (ALL, n?=?101) and healthy controls (n?=?471). Association between SNPs and leukemia risk was calculated by estimating odds ratios in the multivariate logistic regression analysis. For miRSNPs that were associated with leukemia risk we performed luciferase reporter assays to examine whether they influence miRNA binding. RESULTS: Here we show that variant alleles of TLX1_rs2742038 and ETV6_rs1573613 were associated with increased risk of childhood ALL (OR (95% CI)?=?3.97 (1.43-11.02) and 1.9 (1.16-3.11), respectively), while PML_rs9479 was associated with decreased ALL risk (OR?=?0.55 (0.36-0.86). In adult myeloid leukemias we found significant associations between the variant allele of PML_rs9479 and decreased AML risk (OR?=?0.61 (0.38-0.97), and between variant alleles of IRF8_ rs10514611 and ARHGAP26_rs187729 and increased CML risk (OR?=?2.4 (1.12-5.15) and 1.63 (1.07-2.47), respectively). Moreover, we observed a significant trend for an increasing ALL and CML risk with the growing number of risk genotypes with OR?=?13.91 (4.38-44.11) for carriers of ?3 risk genotypes in ALL and OR?=?4.9 (1.27-18.85) for carriers of 2 risk genotypes in CML. Luciferase reporter assays revealed that the C allele of ARHGAP26_rs187729 creates an illegitimate binding site for miR-18a-3p, while the A allele of PML_rs9479 enhances binding of miR-510-5p and the C allele of ETV6_rs1573613 weakens binding of miR-34c-5p and miR-449b-5p. CONCLUSIONS: Our study implicates that microRNA-binding site polymorphisms modulate leukemia risk by interfering with the miRNA-mediated regulation. Our findings underscore the significance of variability in 3' untranslated regions in leukemia.

Dataset Information

Uncovering low-dimensional, miR-based signatures of acute myeloid and lymphoblastic leukemias with a machine-learning-driven network approach.

Publications

Uncovering low-dimensional, miR-based signatures of acute myeloid and lymphoblastic leukemias with a machine-learning-driven network approach.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets