A semi-supervised Bayesian mixture modelling approach for joint batch correction and classification

Systematic differences between batches of samples present significant challenges when analysing biological data. Such batch effects are well-studied and are liable to occur in any setting where multiple batches are assayed. Many existing methods for accounting for these have focused on high-dimensional data such as RNA-seq and have assumptions that reflect this. Here we focus on batch-correction in low-dimensional classification problems. We propose a semi-supervised Bayesian generative classifier based on mixture models that jointly predicts class labels and models batch effects. Our model allows observations to be probabilistically assigned to classes in a way that incorporates uncertainty arising from batch effects. We explore two choices for the within-class densities: the multivariate normal and the multivariate t. A simulation study demonstrates that our method performs well compared to popular off-the-shelf machine learning methods and is also quick; performing 15,000 iterations on a dataset of 500 samples with 2 measurements each in 7.3 seconds for the MVN mixture model and 11.9 seconds for the MVT mixture model. We apply our model to two datasets generated using the enzyme-linked immunosorbent assay (ELISA), a spectrophotometric assay often used to screen for antibodies. The examples we consider were collected in 2020 and measure seropositivity for SARS-CoV-2. We use our model to estimate seroprevalence in the populations studied. We implement the models in C++ using a Metropolis-within-Gibbs algorithm; this is available in the R package at https://github.com/stcolema/BatchMixtureModel. Scripts to recreate our analysis are at https://github.com/stcolema/BatchClassifierPaper.

SELINA: Single-cell Assignment using Multiple-Adversarial Domain Adaptation Network with Large-scale References

The rapid accumulation of single-cell RNA-seq data has provided rich resources to characterize various human cell types. Cell type annotation is the critical step in analyzing single-cell RNA-seq data. However, accurate cell type annotation based on public references is challenging due to the inconsistent annotations, batch effects, and poor characterization of rare cell types. Here, we introduce SELINA (single cELl identity NAvigator), an integrative annotation transferring framework for automatic cell type annotation. SELINA optimizes the annotation for minority cell types by synthetic minority over-sampling, removes batch effects among reference datasets using a multiple-adversarial domain adaptation network (MADA), and fits the query data with reference data using an autoencoder. Finally, SELINA affords a comprehensive and uniform reference atlas with 1.7 million cells covering 230 major human cell types. We demonstrated the robustness and superiority of SELINA in most human tissues compared to existing methods. SELINA provided a one-stop solution for human single- cell RNA-seq data annotation with the potential to extend for other species.

GMQN: A Reference-Based Method for Correcting Batch Effects and Probe Bias in HumanMethylation BeadChip

The Illumina HumanMethylation BeadChip is one of the most cost-effective methods to quantify DNA methylation levels at single-base resolution across the human genome, which makes it a routine platform for epigenome-wide association studies. It has accumulated tens of thousands of DNA methylation array samples in public databases, providing great support for data integration and further analysis. However, the majority of public DNA methylation data are deposited as processed data without background probes which are widely used in data normalization. Here, we present Gaussian mixture quantile normalization (GMQN), a reference based method for correcting batch effects as well as probe bias in the HumanMethylation BeadChip. Availability and implementation: https://github.com/MengweiLi-project/gmqn.

Benchmarking atlas-level data integration in single-cell genomics

Single-cell atlases often include samples that span locations, laboratories and conditions, leading to complex, nested batch effects in data. Thus, joint analysis of atlas datasets requires reliable data integration. To guide integration method choice, we benchmarked 68 method and preprocessing combinations on 85 batches of gene expression, chromatin accessibility and simulation data from 23 publications, altogether representing >1.2 million cells distributed in 13 atlas-level integration tasks. We evaluated methods according to scalability, usability and their ability to remove batch effects while retaining biological variation using 14 evaluation metrics. We show that highly variable gene selection improves the performance of data integration methods, whereas scaling pushes methods to prioritize batch removal over conservation of biological variation. Overall, scANVI, Scanorama, scVI and scGen perform well, particularly on complex integration tasks, while single-cell ATAC-sequencing integration performance is strongly affected by choice of feature space. Our freely available Python module and benchmarking pipeline can identify optimal data integration methods for new data, benchmark new methods and improve method development.

Extent, impact, and mitigation of batch effects in tumor biomarker studies using tissue microarrays

Tissue microarrays (TMAs) have been used in thousands of cancer biomarker studies. To what extent batch effects, measurement error in biomarker levels between slides, affects TMA-based studies has not been assessed systematically. We evaluated 20 protein biomarkers on 14 TMAs with prospectively collected tumor tissue from 1,448 primary prostate cancers. In half of the biomarkers, more than 10% of biomarker variance was attributable to between-TMA differences (range, 1-48%). We implemented different methods to mitigate batch effects (R package batchtma), tested in plasmode simulation. Biomarker levels were more similar between mitigation approaches compared to uncorrected values. For some biomarkers, associations with clinical features changed substantially after addressing batch effects. Batch effects and resulting bias are not an error of an individual study but an inherent feature of TMA-based protein biomarker studies. They always need to be considered during study design and addressed analytically in studies using more than one TMA.

Batch effect reduction of microarray data with dependent samples using an empirical Bayes approach (BRIDGE)

Batch-effects present challenges in the analysis of high-throughput molecular data and are particularly problematic in longitudinal studies when interest lies in identifying genes/features whose expression changes over time, but time is confounded with batch. While many methods to correct for batch-effects exist, most assume independence across samples; an assumption that is unlikely to hold in longitudinal microarray studies. We propose Batch effect Reduction of mIcroarray data with Dependent samples usinG Empirical Bayes (BRIDGE), a three-step parametric empirical Bayes approach that leverages technical replicate samples profiled at multiple timepoints/batches, so-called “bridge samples”, to inform batch-effect reduction/attenuation in longitudinal microarray studies. Extensive simulation studies and an analysis of a real biological data set were conducted to benchmark the performance of BRIDGE against both ComBat and longitudinal ComBat. Our results demonstrate that while all methods perform well in facilitating accurate estimates of time effects, BRIDGE outperforms both ComBat and longitudinal ComBat in the removal of batch-effects in data sets with bridging samples, and perhaps as a result, was observed to have improved statistical power for detecting genes with a time effect. BRIDGE demonstrated competitive performance in batch effect reduction of confounded longitudinal microarray studies, both in simulated and a real data sets, and may serve as a useful preprocessing method for researchers conducting longitudinal microarray studies that include bridging samples.

CIDER: an interpretable meta-clustering framework for single-cell RNA-seq data integration and evaluation

Clustering of joint single-cell RNA-Seq (scRNA-Seq) data is often challenged by confounding factors, such as batch effects and biologically relevant variability. Existing batch effect removal methods typically require strong assumptions on the composition of cell populations being near identical across samples. Here, we present CIDER, a meta-clustering workflow based on inter-group similarity measures. We demonstrate that CIDER outperforms other scRNA-Seq clustering methods and integration approaches in both simulated and real datasets. Moreover, we show that CIDER can be used to assess the biological correctness of integration in real datasets, while it does not require the existence of prior cellular annotations.

xDEEP-MSI: Explainable Bias-Rejecting Microsatellite Instability Deep Learning System in Colorectal Cancer

The prediction of microsatellite instability (MSI) using deep learning (DL) techniques could have significant benefits, including reducing cost and increasing MSI testing of colorectal cancer (CRC) patients. Nonetheless, batch effects or systematic biases are not well characterized in digital histology models and lead to overoptimistic estimates of model performance. Methods to not only palliate but to directly abrogate biases are needed. We present a multiple bias rejecting DL system based on adversarial networks for the prediction of MSI in CRC from tissue microarrays (TMAs), trained and validated in 1788 patients from EPICOLON and HGUA. The system consists of an end-to-end image preprocessing module that tile samples at multiple magnifications and a tissue classification module linked to the bias-rejecting MSI predictor. We detected three biases associated with the learned representations of a baseline model: the project of origin of samples, the patient’s spot and the TMA glass where each spot was placed. The system was trained to directly avoid learning the batch effects of those variables. The learned features from the bias-ablated model achieved maximum discriminative power with respect to the task and minimal statistical mean dependence with the biases. The impact of different magnifications, types of tissues and the model performance at tile vs patient level is analyzed. The AUC at tile level, and including all three selected tissues (tumor epithelium, mucin and lymphocytic regions) and 4 magnifications, was 0.87 ± 0.03 and increased to 0.9 ± 0.03 at patient level. To the best of our knowledge, this is the first work that incorporates a multiple bias ablation technique at the DL architecture in digital pathology, and the first using TMAs for the MSI prediction task.

PMD Uncovers Widespread Cell-State Erasure by scRNAseq Batch Correction Methods

Single cell RNAseq (scRNAseq) batches range from technical replicates to multi-tissue atlases, thus requiring robust batch correction methods that operate effectively across this similarity spectrum. Currently, no metrics allow for full benchmarking across this spectrum, resulting in benchmarks that quantify removal of batch effects without quantifying preservation of real batch differences. Here, we address these gaps with a new statistical metric [Percent Maximum Difference (PMD)] that linearly quantifies batch similarity, and simulations generating cells from mixtures of distinct gene expression programs (cell-lineages/-types/-states). Using 690 real-world and 672 simulated integrations (7.2e6 cells total) we compared 7 batch integration approaches across the spectrum of similarity with batch-confounded gene expression. Count downsampling appeared the most robust, while others left residual batch effects or produced over-merged datasets. We further released open-source PMD and downsampling packages, with the latter capable of downsampling an organism atlas (245,389 cells) in tens of minutes on a standard computer.