Background: Detection of low abundance metabolites is important for de novo mapping of metabolic pathways related to diet, microbiome or environmental exposures. Multiple algorithms are available to extract m/z features from liquid chromatography-mass spectral data in a conservative manner, which tends to preclude detection of low abundance chemicals and chemicals found in small subsets of samples. The present study provides software to enhance such algorithms for feature detection, quality assessment, and annotation. Results: xMSanalyzer is a set of utilities for automated processing of metabolomics data. The utilites can be classified into four main modules to: 1) improve feature detection for replicate analyses by systematic re-extraction with multiple parameter settings and data merger to optimize the balance between sensitivity and reliability, 2) evaluate sample quality and feature consistency, 3) detect feature overlap between datasets, and 4) characterize high-resolution m/z matches to small molecule metabolites and biological pathways using multiple chemical databases. The package was tested with plasma samples and shown to more than double the number of features extracted while improving quantitative reliability of detection. MS/MS analysis of a random subset of peaks that were exclusively detected using xMSanalyzer confirmed that the optimization scheme improves detection of real metabolites. Conclusions: xMSanalyzer is a package of utilities for data extraction, quality control assessment, detection of overlapping and unique metabolites in multiple datasets, and batch annotation of metabolites. The program was designed to integrate with existing packages such as apLCMS and XCMS, but the framework can also be used to enhance data extraction for other LC/MS data software.

In predictive model development, gene expression data is associated with the unique challenge that the number of samples (n) is much smaller than the amount of features ( p ). This “ n ≪ p ” property has prevented classification of gene expression data from deep learning techniques, which have been proved powerful under “ n > p ” scenarios in other application fields, such as image classification. Further, the sparsity of effective features with unknown correlation structures in gene expression profiles brings more challenges for classification tasks. To tackle these problems, we propose a newly developed classifier named Forest Deep Neural Network (fDNN), to integrate the deep neural network architecture with a supervised forest feature detector. Using this built-in feature detector, the method is able to learn sparse feature representations and feed the representations into a neural network to mitigate the overfitting problem. Simulation experiments and real data analyses using two RNA-seq expression datasets are conducted to evaluate fDNN’s capability. The method is demonstrated a useful addition to current predictive models with better classification performance and more meaningful selected features compared to ordinary random forests and deep neural networks.

Gene expression data represents a unique challenge in predictive model building, because of the small number of samples (n) compared with the huge amount of features (p). This ‘$n≪p$’ property has hampered application of deep learning techniques for disease outcome classification. Sparse learning by incorporating external gene network information could be a potential solution to this issue. Still, the problem is very challenging because (i) there are tens of thousands of features and only hundreds of training samples, (ii) the scale-free structure of the gene network is unfriendly to the setup of convolutional neural networks. To address these issues and build a robust classification model, we propose the Graph-Embedded Deep Feedforward Networks (GEDFN), to integrate external relational information of features into the deep neural network architecture. The method is able to achieve sparse connection between network layers to prevent overfitting. To validate the method’s capability, we conducted both simulation experiments and real data analysis using a breast invasive carcinoma RNA-seq dataset and a kidney renal clear cell carcinoma RNA-seq dataset from The Cancer Genome Atlas. The resulting high classification accuracy and easily interpretable feature selection results suggest the method is a useful addition to the current graph-guided classification models and feature selection procedures. The method is available at https://github.com/yunchuankong/GEDFN. Supplementary data are available at Bioinformatics online.

With the growth of metabolomics research, more and more studies are conducted on large numbers of samples. Due to technical limitations of the Liquid Chromatography–Mass Spectrometry (LC/MS) platform, samples often need to be processed in multiple batches. Across different batches, we often observe differences in data characteristics. In this work, we specifically focus on data generated in multiple batches on the same LC/MS machinery. Traditional preprocessing methods treat all samples as a single group. Such practice can result in errors in the alignment of peaks, which cannot be corrected by post hoc application of batch effect correction methods. In this work, we developed a new approach that address the batch effect issue in the preprocessing stage, resulting in better peak detection, alignment and quantification. It can be combined with down-stream batch effect correction methods to further correct for between-batch intensity differences. The method is implemented in the existing workflow of the apLCMS platform. Analyzing data with multiple batches, both generated from standardized quality control (QC) plasma samples and from real biological studies, the new method resulted in feature tables with better consistency, as well as better down-stream analysis results. The method can be a useful addition to the tools available for large studies involving multiple batches. The method is available as part of the apLCMS package. Download link and instructions are athttps://mypage.cuhk.edu.cn/academics/yutianwei/apLCMS/.

Cell clustering is one of the most important and commonly performed tasks in single-cell RNA sequencing (scRNA-seq) data analysis. An important step in cell clustering is to select a subset of genes (referred to as ‘features’), whose expression patterns will then be used for downstream clustering. A good set of features should include the ones that distinguish different cell types, and the quality of such set could have a significant impact on the clustering accuracy. All existing scRNA-seq clustering tools include a feature selection step relying on some simple unsupervised feature selection methods, mostly based on the statistical moments of gene-wise expression distributions. In this work, we carefully evaluate the impact of feature selection on cell clustering accuracy. In addition, we develop a feature selection algorithm named FEAture SelecTion (FEAST), which provides more representative features. We apply the method on 12 public scRNA-seq datasets and demonstrate that using features selected by FEAST with existing clustering tools significantly improve the clustering accuracy.