Dataset Information

Cross-platform normalization enables machine learning model training on microarray and RNA-seq data simultaneously.

ABSTRACT: Large compendia of gene expression data have proven valuable for the discovery of novel biological relationships. Historically, most available RNA assays were run on microarray, while RNA-seq is now the platform of choice for many new experiments. The data structure and distributions between the platforms differ, making it challenging to combine them directly. Here we perform supervised and unsupervised machine learning evaluations to assess which existing normalization methods are best suited for combining microarray and RNA-seq data. We find that quantile and Training Distribution Matching normalization allow for supervised and unsupervised model training on microarray and RNA-seq data simultaneously. Nonparanormal normalization and z-scores are also appropriate for some applications, including pathway analysis with Pathway-Level Information Extractor (PLIER). We demonstrate that it is possible to perform effective cross-platform normalization using existing methods to combine microarray and RNA-seq data for machine learning applications.

SUBMITTER: Foltz SM

PROVIDER: S-EPMC9968332 | biostudies-literature | 2023 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Cross-platform normalization enables machine learning model training on microarray and RNA-seq data simultaneously.

Foltz Steven M SM Greene Casey S CS Taroni Jaclyn N JN

Communications biology 20230225 1

Large compendia of gene expression data have proven valuable for the discovery of novel biological relationships. Historically, most available RNA assays were run on microarray, while RNA-seq is now the platform of choice for many new experiments. The data structure and distributions between the platforms differ, making it challenging to combine them directly. Here we perform supervised and unsupervised machine learning evaluations to assess which existing normalization methods are best suited f ...[more]

PMID: 36841852

Similar Datasets

Project description:Acute promyelocytic leukaemia (APL) can be cured by the co-administration of arsenic trioxide (ATO) and all-trans retinoic acid (ATRA). These small molecules relieve the differentiation blockade of the transformed promyelocytes and trigger their maturation into functional neutrophils, which are physiologically primed for apoptosis. This normalization therapy represents a compelling alternative to cytotoxic anticancer chemotherapy, but lacks an in vitro model system for testing the efficiency of novel combination treatments consisting of inducers of differentiation and metallopharmaceuticals. Here, using proteome profiling we present an experimental framework that enables characterising the differentiation- and metal-specific effects of the combination treatment in a panel of acute myeloid leukaemia (AML) cell lines (HL-60 and U937), including APL (NB4). Differentiation had a substantial impact on the proteome on the order of 10% of the identified proteins and featured classical markers and transcription factors of myeloid differentiation. Additionally, ATO provoked specific cytoprotective effects in the AML cell lines HL-60 and U937. In HL-60, these effects included an integrated stress response (ISR) in conjunction with redox defence, while proteasomal responses and a metabolic rewiring were observed in U937 cells. In contrast, the APL cell line NB4 did not display such adaptions indicating a lack of plasticity to cope with the metal-induced stress, which may explain the clinical success of this combination treatment. Based on the induction of these cytoprotective effects, we proposed a novel metal-based compound to be used for the combination treatment instead of ATO. The organoruthenium drug candidate plecstatin-1 was previously shown to induce reactive oxygen species and an ISR. Indeed, the plecstatin-1 combination was found to affect similar pathways compared to the ATO combination in HL-60 cells and did not lead to cytoprotective response signatures in NB4. Moreover, the monocytic cell line U937 showed a low plasticity to cope with the plecstatin-1 combination, which suggests that this combination might achieve therapeutic benefit beyond APL. We propose that the cytoprotective plasticity of cancer cells might serve as a general proxy to discover novel combination treatments in vitro.

Dataset Information

Cross-platform normalization enables machine learning model training on microarray and RNA-seq data simultaneously.

Publications

Cross-platform normalization enables machine learning model training on microarray and RNA-seq data simultaneously.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets