Search bioRxivSearch

Biology subjects

Varshney, L. R.

Publications and source records attributed to Varshney, L. R..

2 recordsLinked to original sources

Optimal recovery of missing values for non-negative matrix factorization

We extend the approximation-theoretic technique of optimal recovery to the setting of imputing missing values in clustered data, specifically for non-negative matrix factorization (NMF), and develop an implementable algorithm. Under certain geometric conditions, we prove tight upper bounds on NMF relative error, which is the first bound of this type for missing values. We also give probabilistic bounds for the same geometric assumptions. Experiments on image data and biological data show that this theoretically-grounded technique performs as well as or better than other imputation techniques that account for local structure.

bioinformatics

Registration for Image-Based Transcriptomics: Parametric Signal Features and Multivariate Information Measures

Image-based transcriptomics involves determining spatial patterns in gene expression across cells and tissues. Image registration is a necessary component of data analysis pipelines that study gene expression levels across different cells and intracellular structures. We consider images from multiplexed single molecule fluorescent in situ hybridization (smFISH) and multiplexed in situ sequencing (ISS) datasets from the Human Cell Atlas project and demonstrate a novel approach to groupwise image registration using a parametric representation of images based on finite rate of innovation sampling, together with practical optimization of empirical multivariate information measures.

bioinformatics