Haiyan Huang

Pharmacogenomics studies are attracting an increasing amount of interest from researchers in precision medicine. The advances in high-throughput experiments and multiplexed approaches allow the large-scale quantification of drug sensitivities in molecularly characterized cancer cell lines (CCLs), resulting in a number of open drug sensitivity datasets for drug biomarker discovery. However, a significant inconsistency in drug sensitivity values among these datasets has been noted. Such inconsistency indicates the presence of substantial noise, subsequently hindering downstream analyses. To address the noise in drug sensitivity data, we introduce a robust and scalable deep learning framework, Residual Thresholded Deep Matrix Factorization (RT-DMF). This method takes a single drug sensitivity data matrix as its sole input and outputs a corrected and imputed matrix. Deep matrix factorization (DMF) excels at uncovering subtle patterns, due to its minimal reliance on data structure assumptions. This attribute significantly boosts DMF’s ability to identify complex hidden patterns among nuisance effects in the data, thereby facilitating the detection of signals that are therapeutically relevant. Furthermore, RT-DMF incorporates an iterative residual thresholding procedure, which plays a crucial role in retaining signals more likely to hold therapeutic importance. Validation using simulated datasets and real pharmacogenomics datasets demonstrates the effectiveness of our approach in correcting noise and imputing missing data in drug sensitivity datasets (open-source package available at https://github.com/tomwhoooo/rtdmf).

Multiple sclerosis is a chronic autoimmune disease that affects the central nervous system. Understanding multiple sclerosis progression and identifying the implicated brain structures is crucial for personalized treatment decisions. Deformation-based morphometry utilizes anatomical magnetic resonance imaging to quantitatively assess volumetric brain changes at the voxel level, providing insight into how each brain region contributes to clinical progression with regards to neurodegeneration. Utilizing such voxel-level data from a relapsing multiple sclerosis clinical trial, we extend a model-agnostic feature importance metric to identify a robust and predictive feature set that corresponds to clinical progression. These features correspond to brain regions that are clinically meaningful in MS disease research, demonstrating their scientific relevance. When used to predict progression using classical survival models and 3D convolutional neural networks, the identified regions led to the best-performing models, demonstrating their prognostic strength. We also find that these features generalize well to other definitions of clinical progression and can compensate for the omission of highly prognostic clinical features, underscoring the predictive power and clinical relevance of deformation-based morphometry as a regional identification tool.

A classical result for the simple symmetric random walk with 2nsteps is that the number of steps above the origin, the time of the last visit to the origin, and the time of the maximum height all have exactly the same distribution and converge when scaled to the arcsine law. Motivated by applications in genomics, we study the distributions of these statistics for the non-Markovian random walk generated from the ascents and descents of a uniform random permutation and a Mallows(q) permutation and show that they have the same asymptotic distributions as for the simple random walk. We also give an unexpected conjecture, along with numerical evidence and a partial proof in special cases, for the result that the number of steps above the origin by step 2n for the uniform permutation generated walk has exactly the same discrete arcsine distribution as for the simple random walk, even though the other statistics for these walks have very different laws. We also give explicit error bounds to the limit theorems using Stein’s method for the arcsine distribution, as well as functional central limit theorems and a strong embedding of the Mallows(q) permutation which is of independent interest.

In this article we propose a novel ranking algorithm, referred to as HierLPR, for the multi-label classification problem when the candidate labels follow a known hierarchical structure. HierLPR is motivated by a new metric called eAUC that we design to assess the ranking of classification decisions. This metric, associated with the hit curve and local precision rate, emphasizes the accuracy of the first calls. We show that HierLPR optimizes eAUC under the tree constraint and some light assumptions on the dependency between the nodes in the hierarchy. We also provide a strategy to make calls for each node based on the ordering produced by HierLPR, with the intent of controlling FDR or maximizing F-score. The performance of our proposed methods is demonstrated on synthetic datasets as well as a real example of disease diagnosis using NCBI GEO datasets. In these cases, HierLPR shows a favorable result over competing methods in the early part of the precision-recall curve.

Background: Developing appropriate computational tools to distill biological insights from large‐scale gene expression data has been an important part of systems biology. Considering that gene relationships may change or only exist in a subset of collected samples, biclustering that involves clustering both genes and samples has become in‐creasingly important, especially when the samples are pooled from a wide range of experimental conditions.Methods: In this paper, we introduce a new biclustering algorithm to find subsets of genomic expression features (EFs) (e.g., genes, isoforms, exon inclusion) that show strong “group interactions” under certain subsets of samples. Group interactions are defined by strong partial correlations, or equivalently, conditional dependencies between EFs after removing the influences of a set of other functionally related EFs. Our new biclustering method, named SCCA‐BC, extends an existing method for group interaction inference, which is based on sparse canonical correlation analysis (SCCA) coupled with repeated random partitioning of the gene expression data set.Results: SCCA‐BC gives sensible results on real data sets and outperforms most existing methods in simulations. Software is available at https://github.com/pimentel/scca‐bc.Conclusions: SCCA‐BC seems to work in numerous conditions and the results seem promising for future extensions. SCCA‐BC has the ability to find different types of bicluster patterns, and it is especially advantageous in identifying a bicluster whose elements share the same progressive and multivariate normal distribution with a dense covariance matrix.

Capturing association patterns in gene expression levels under different conditions or time points is important for inferring gene regulatory interactions. In practice, temporal changes in gene expression may result in complex association patterns that require more sophisticated detection methods than simple correlation measures. For instance, the effect of regulation may lead to time-lagged associations and interactions local to a subset of samples. Furthermore, expression profiles of interest may not be aligned or directly comparable (e.g. gene expression profiles from two species). We propose a count statistic for measuring association between pairs of gene expression profiles consisting of ordered samples (e.g. time-course), where correlation may only exist locally in subsequences separated by a position shift. The statistic is simple and fast to compute, and we illustrate its use in two applications. In a cross-species comparison of developmental gene expression levels, we show our method not only measures association of gene expressions between the two species, but also provides alignment between different developmental stages. In the second application, we applied our statistic to expression profiles from two distinct phenotypic conditions, where the samples in each profile are ordered by the associated phenotypic values. The detected associations can be useful in building correspondence between gene association networks under different phenotypes. On the theoretical side, we provide asymptotic distributions of the statistic for different regions of the parameter space and test its power on simulated data. The code used to perform the analysis is available as part of the Supplementary Material . Supplementary data are available at Bioinformatics online.

One goal of single-cell RNA sequencing (scRNA seq) is to expose possible heterogeneity within cell populations due to meaningful, biological variation. Examining cell-to-cell heterogeneity, and further, identifying subpopulations of cells based on scRNA seq data has been of common interest in life science research. A key component to successfully identifying cell subpopulations (or clustering cells) is the (dis)similarity measure used to group the cells. In this paper, we introduce a novel measure, named SIDEseq, to assess cell-to-cell similarity using scRNA seq data. SIDEseq first identifies a list of putative differentially expressed (DE) genes for each pair of cells. SIDEseq then integrates the information from all the DE gene lists (corresponding to all pairs of cells) to build a similarity measure between two cells. SIDEseq can be implemented in any clustering algorithm that requires a (dis)similarity matrix. This new measure incorporates information from all cells when evaluating the similarity between any two cells, a characteristic not commonly found in existing (dis)similarity measures. This property is advantageous for two reasons: (a) borrowing information from cells of different subpopulations allows for the investigation of pairwise cell relationships from a global perspective and (b) information from other cells of the same subpopulation could help to ensure a robust relationship assessment. We applied SIDEseq to a newly generated human ovarian cancer scRNA seq dataset, a public human embryo scRNA seq dataset, and several simulated datasets. The clustering results suggest that the SIDEseq measure is capable of uncovering important relationships between cells, and outperforms or at least does as well as several popular (dis)similarity measures when used on these datasets.

Networks pervade many disciplines of science for analyzing complex systems with interacting components. In particular, this concept is commonly used to model interactions between genes and identify closely associated genes forming functional modules. In this paper, we focus on gene group interactions and infer these interactions using appropriate partial correlations between genes, that is, the conditional dependencies between genes after removing the influences of a set of other functionally related genes. We introduce a new method for estimating group interactions using sparse canonical correlation analysis (SCCA) coupled with repeated random partition and subsampling of the gene expression data set. By considering different subsets of genes and ways of grouping them, our interaction measure can be viewed as an aggregated estimate of partial correlations of different orders. Our approach is unique in evaluating conditional dependencies when the correct dependent sets are unknown or only partially known. As a result, a gene network can be constructed using the interaction measures as edge weights and gene functional groups can be inferred as tightly connected communities from the network. Comparisons with several popular approaches using simulated and real data show our procedure improves both the statistical significance and biological interpretability of the results. In addition to achieving considerably lower false positive rates, our procedure shows better performance in detecting important biological pathways.