BERMUDA: a novel deep transfer learning method for single-cell RNA sequencing batch correction reveals hidden high-resolution cellular subtypes.
Ontology highlight
ABSTRACT:
SUBMITTER: Wang T
PROVIDER: S-EPMC6691531 | biostudies-literature | 2019 Aug
REPOSITORIES: biostudies-literature
Publications
BERMUDA: a novel deep transfer learning method for single-cell RNA sequencing batch correction reveals hidden high-resolution cellular subtypes.
Wang Tongxin T Johnson Travis S TS Shao Wei W Lu Zixiao Z Helm Bryan R BR Zhang Jie J Huang Kun K
Genome biology 20190812 1
To fully utilize the power of single-cell RNA sequencing (scRNA-seq) technologies for identifying cell lineages and bona fide transcriptional signals, it is necessary to combine data from multiple experiments. We present BERMUDA (Batch Effect ReMoval Using Deep Autoencoders), a novel transfer-learning-based method for batch effect correction in scRNA-seq data. BERMUDA effectively combines different batches of scRNA-seq data with vastly different cell population compositions and amplifies biologi ...[more]
Similar Datasets
Project description:We validated the use of a novel image-based method for partial volume correction (PVC), structural-functional synergistic resolution recovery (SFS-RR) for the accurate quantification of dopamine synthesis capacity measured using [(18)F]DOPA positron emission tomography. The bias and reliability of SFS-RR were compared with the geometric transfer matrix (GTM) method. Both methodologies were applied to the parametric maps of [(18)F]DOPA utilization rates (ki(cer)). Validation was first performed by measuring repeatability on test-retest scans. The precision of the methodologies instead was quantified using simulated [(18)F]DOPA images. The sensitivity to the misspecification of the full-width-half-maximum (FWHM) of the scanner point-spread-function on both approaches was also assessed. In the in-vivo data, the ki(cer) was significantly increased by application of both PVC procedures while the reliability remained high (intraclass correlation coefficients >0.85). The variability was not significantly affected by either PVC approach (<10% variability in both cases). The corrected ki(cer) was significantly influenced by the FWHM applied in both the acquired and simulated data. This study shows that SFS-RR can effectively correct for partial volume effects to a comparable degree to GTM but with the added advantage that it enables voxelwise analyses, and that the FWHM used can affect the PVC result indicating the importance of accurately calibrating the FWHM used in the recovery model.
| S-EPMC3677111 | biostudies-other
Project description:BackgroundAn increasing number of studies within digital pathology show the potential of artificial intelligence (AI) to diagnose cancer using histological whole slide images, which requires large and diverse data sets. While diversification may result in more generalizable AI-based systems, it can also introduce hidden variables. If neural networks are able to distinguish/learn hidden variables, these variables can introduce batch effects that compromise the accuracy of classification systems.ObjectiveThe objective of the study was to analyze the learnability of an exemplary selection of hidden variables (patient age, slide preparation date, slide origin, and scanner type) that are commonly found in whole slide image data sets in digital pathology and could create batch effects.MethodsWe trained four separate convolutional neural networks (CNNs) to learn four variables using a data set of digitized whole slide melanoma images from five different institutes. For robustness, each CNN training and evaluation run was repeated multiple times, and a variable was only considered learnable if the lower bound of the 95% confidence interval of its mean balanced accuracy was above 50.0%.ResultsA mean balanced accuracy above 50.0% was achieved for all four tasks, even when considering the lower bound of the 95% confidence interval. Performance between tasks showed wide variation, ranging from 56.1% (slide preparation date) to 100% (slide origin).ConclusionsBecause all of the analyzed hidden variables are learnable, they have the potential to create batch effects in dermatopathology data sets, which negatively affect AI-based classification systems. Practitioners should be aware of these and similar pitfalls when developing and evaluating such systems and address these and potentially other batch effect variables in their data sets through sufficient data set stratification.
| S-EPMC7886613 | biostudies-literature
Project description:Recent developments of single-cell RNA-seq (scRNA-seq) technologies have led to enormous biological discoveries. As the scale of scRNA-seq studies increases, a major challenge in analysis is batch effects, which are inevitable in studies involving human tissues. Most existing methods remove batch effects in a low-dimensional embedding space. Although useful for clustering, batch effects are still present in the gene expression space, leaving downstream gene-level analysis susceptible to batch effects. Recent studies have shown that batch effect correction in the gene expression space is much harder than in the embedding space. Methods such as Seurat 3.0 rely on the mutual nearest neighbor (MNN) approach to remove batch effects in gene expression, but MNN can only analyze two batches at a time, and it becomes computationally infeasible when the number of batches is large. Here, we present CarDEC, a joint deep learning model that simultaneously clusters and denoises scRNA-seq data while correcting batch effects both in the embedding and the gene expression space. Comprehensive evaluations spanning different species and tissues showed that CarDEC outperforms Scanorama, DCA + Combat, scVI, and MNN. With CarDEC denoising, non-highly variable genes offer as much signal for clustering as the highly variable genes (HVGs), suggesting that CarDEC substantially boosted information content in scRNA-seq. We also showed that trajectory analysis using CarDEC's denoised and batch-corrected expression as input revealed marker genes and transcription factors that are otherwise obscured in the presence of batch effects. CarDEC is computationally fast, making it a desirable tool for large-scale scRNA-seq studies.
| S-EPMC8494213 | biostudies-literature
Project description:Motivation: Sample source, procurement process, and other technical variations introduce batch effects into genomics data. Algorithms to remove these artifacts enhance differences between known biological covariates, but also carry potential concern of removing intra-group biological heterogeneity and thus any personalized genomic signatures. As a result, accurate identification of novel subtypes from batch corrected genomics data is challenging using standard algorithms designed to remove batch effects for class comparison analyses. Nor can batch effects be corrected reliably in future applications of genomics-based clinical tests, in which the biological groups are by definition unknown a priori. Results: Therefore, we introduce new algorithm, personalized-SVA (pSVA), blind to biological covariates corrected technical artifacts while retaining biological heterogeneity in genomic data. This algorithm facilitated accurate subtype identification in head and neck cancer from gene expression data in both formalin fixed and frozen samples. When applied to predict HPV status, pSVA improved cross- study validation even if the sample batches were highly confounded with HPV status in the training set. Availability: All analyses were performed using R version 2.15.0. The code and data used to generate the results of this manuscript is available from https://sourceforge.net/projects/psva.
2016-12-12 | GSE53355 | GEO
Project description:Microarray batch effect (BE) has been the primary bottleneck for large-scale integration of data from multiple experiments. Current BE correction methods either need known batch identities (ComBat) or have the potential to overcorrect, by removing true but unknown biological differences (Surrogate Variable Analysis SVA). It is well known that experimental conditions such as array or reagent batches, PCR amplification or ozone levels can affect the measured expression levels; often the direction of perturbation of the measured expression is the same in different datasets. However, there are no BE correction algorithms that attempt to estimate the individual effects of technical differences and use them to correct expression data. In this manuscript, we show that a set of signatures, each of which is a vector the length of the number of probes, calculated on a reference set of microarray samples can predict much of the batch effect in other validation sets. We present a rationale of selecting a reference set of samples designed to estimate technical differences without removing biological differences. Putting both together, we introduce the Batch Effect Signature Correction (BESC) algorithm that uses the BES calculated on the reference set to efficiently predict and remove BE. Using two independent validation sets, we show that BESC is capable of removing batch effect without removing unknown but true biological differences. Much of the variations due to batch effect is shared between different microarray datasets. That shared information can be used to predict signatures (i.e. directions of perturbation) due to batch effect in new datasets. The correction can be precomputed without using the samples to be corrected (blind), done on each sample individually (single sample) and corrects only known technical effects without removing known or unknown biological differences (conservative). Those three characteristics make it ideal for high-throughput correction of samples for a microarray data repository. We also compare the performance of BESC to three other batch correction methods: SVA, Removing Unwanted Variation (RUV) and Hidden Covariates with Prior (HCP). An R Package besc implementing the algorithm is available from http://explainbio.com.
| S-EPMC7145015 | biostudies-literature
Project description:High-dimensional (HD) NMR spectra have poorer digital resolution than low-dimensional (LD) spectra, for a fixed amount of experiment time. This has led to "reduced-dimensionality" strategies, in which several LD projections of the HD NMR spectrum are acquired, each with higher digital resolution; an approximate HD spectrum is then inferred by some means. We propose a strategy that moves in the opposite direction, by adding more time dimensions to increase the information content of the data set, even if only a very sparse time grid is used in each dimension. The full HD time-domain data can be analyzed by the filter diagonalization method (FDM), yielding very narrow resonances along all of the frequency axes, even those with sparse sampling. Integrating over the added dimensions of HD FDM NMR spectra reconstitutes LD spectra with enhanced resolution, often more quickly than direct acquisition of the LD spectrum with a larger number of grid points in each of the fewer dimensions. If the extra-dimensions do not appear in the final spectrum, and are used solely to boost information content, we propose the moniker hidden-dimension NMR. This work shows that HD peaks have unmistakable frequency signatures that can be detected as single HD objects by an appropriate algorithm, even though their patterns would be tricky for a human operator to visualize or recognize, and even if digital resolution in an HD FT spectrum is very coarse compared with natural line widths.
| S-EPMC2659618 | biostudies-literature
Project description:Quality control of biopharmaceuticals such as monoclonal antibodies (mAbs) has been evolving and becoming more challenging as the requirements of the regulatory agencies increase due to the demanding complexity of products under evaluation. Mass Spectrometry (MS)-based methods such as the multi-attribute method (MAM) are being explored to achieve a deeper understanding of the attributes critical for the safety, efficacy, and quality of these products. MAM uses high mass accuracy/high-resolution MS data that enables the direct and simultaneous monitoring of relevant product quality attributes (PQAs, in particular, chemical modifications) in a single workflow, replacing several orthogonal methods, reducing time and costs associated with these assays. Here we describe a MAM implementation process using a QTOF high resolution platform. Method implementation was accomplished using NIST (National Institute for Standards and Technology) mAb reference material and an in-process mAb sample. PQAs as glycosylation profiles, methionine oxidation, tryptophan dioxidation, asparagine deamidation, pyro-Glu at N-terminal and glycation were monitored. Focusing on applications that require batch analysis and high-throughput, sample preparation and LC-MS parameters troubleshooting are discussed. This MAM workflow was successfully explored as reference analytical tool for comprehensive characterization of a downstream processing (DSP) polishing platform and for a comparability study following technology transfer between different laboratories.
| S-EPMC8794205 | biostudies-literature
Project description:Liquid chromatography-mass spectrometry (LC-MS) is a commonly used technique in untargeted metabolomics owing to broad coverage of metabolites, high sensitivity and simple sample preparation. However, data generated from multiple batches are affected by measurement errors inherent to alterations in signal intensity, drift in mass accuracy and retention times between samples both within and between batches. These measurement errors reduce repeatability and reproducibility and may thus decrease the power to detect biological responses and obscure interpretation.Our aim was to develop procedures to address and correct for within- and between-batch variability in processing multiple-batch untargeted LC-MS metabolomics data to increase their quality.Algorithms were developed for: (i) alignment and merging of features that are systematically misaligned between batches, through aggregating feature presence/missingness on batch level and combining similar features orthogonally present between batches; and (ii) within-batch drift correction using a cluster-based approach that allows multiple drift patterns within batch. Furthermore, a heuristic criterion was developed for the feature-wise choice of reference-based or population-based between-batch normalisation.In authentic data, between-batch alignment resulted in picking 15 % more features and deconvoluting 15 % of features previously erroneously aligned. Within-batch correction provided a decrease in median quality control feature coefficient of variation from 20.5 to 15.1 %. Algorithms are open source and available as an R package ('batchCorr').The developed procedures provide unbiased measures of improved data quality, with implications for improved data analysis. Although developed for LC-MS based metabolomics, these methods are generic and can be applied to other data suffering from similar limitations.
| S-EPMC5031781 | biostudies-literature
Project description:With the surge of interest in metabolism and the appreciation of its diverse roles in numerous biomedical contexts, the number of metabolomics studies using liquid chromatography coupled to mass spectrometry (LC-MS) approaches has increased dramatically in recent years. However, variation that occurs independently of biological signal and noise (i.e. batch effects) in metabolomics data can be substantial. Standard protocols for data normalization that allow for cross-study comparisons are lacking. Here, we investigate a number of algorithms for batch effect correction and differential abundance analysis, and compare their performance. We show that linear mixed effects models, which account for latent (i.e. not directly measurable) factors, produce satisfactory results in the presence of batch effects without the need for internal controls or prior knowledge about the nature and sources of unwanted variation in metabolomics data. We further introduce an algorithm-RRmix-within the family of latent factor models and illustrate its suitability for differential abundance analysis in the presence of strong batch effects. Together this analysis provides a framework for systematically standardizing metabolomics data.
| S-EPMC5491020 | biostudies-literature
Project description:MotivationBatch effects are one of the major source of technical variations that affect the measurements in high-throughput studies such as RNA sequencing. It has been well established that batch effects can be caused by different experimental platforms, laboratory conditions, different sources of samples and personnel differences. These differences can confound the outcomes of interest and lead to spurious results. A critical input for batch correction algorithms is the knowledge of batch factors, which in many cases are unknown or inaccurate. Hence, the primary motivation of our paper is to detect hidden batch factors that can be used in standard techniques to accurately capture the relationship between gene expression and other modeled variables of interest.ResultsWe introduce a new algorithm based on data-adaptive shrinkage and semi-Non-negative Matrix Factorization for the detection of unknown batch effects. We test our algorithm on three different datasets: (i) Sequencing Quality Control, (ii) Topotecan RNA-Seq and (iii) Single-cell RNA sequencing (scRNA-Seq) on Glioblastoma Multiforme. We have demonstrated a superior performance in identifying hidden batch effects as compared to existing algorithms for batch detection in all three datasets. In the Topotecan study, we were able to identify a new batch factor that has been missed by the original study, leading to under-representation of differentially expressed genes. For scRNA-Seq, we demonstrated the power of our method in detecting subtle batch effects.Availability and implementationDASC R package is available via Bioconductor or at https://github.com/zhanglabNKU/DASC.Contactzhanghan@nankai.edu.cn or zhandonl@bcm.edu.Supplementary informationSupplementary data are available at Bioinformatics online.
| S-EPMC6454417 | biostudies-literature