Nonnegative Matrix Factorization Model-Based Construction For Molecular Clustering and Prognostic Assessment of Squamous Carcinoma of Head and Neck

Purpose: Exploring nonnegative matrix factorization (NMF) model-based clustering and prognostic modeling of head and neck squamous carcinoma (HNSCC). Methods: The transcriptome microarray data of HNSCC samples were downloaded from The Cancer Genome Atlas (TCGA) and Shanghai Ninth People’s Hospital, and NMF clustering was constructed using the R software package. Relevant prognostic models were developed based on clustering. Results: Based on NMF, all samples were divided into 2 subgroups. Predictive models were constructed by analysing the differential gene between the two subgroups. Results of survival analysis in the current study revealed that the high-risk group had a poor prognosis. Further, results of multi-factor Cox regression analysis revealed that the predictive model was an independent predictor of prognosis. Conclusion: It was evident that the NMF-based prognostic model is a useful guide to the prognostic assessment of HNSCC.

Hybrid Clustering of Single-Cell Gene Expression and Spatial Information via Integrated NMF and K-Means

Advances in single cell transcriptomics have allowed us to study the identity of single cells. This has led to the discovery of new cell types and high resolution tissue maps of them. Technologies that measure multiple modalities of such data add more detail, but they also complicate data integration. We offer an integrated analysis of the spatial location and gene expression profiles of cells to determine their identity. We propose scHybridNMF (single-cell Hybrid Nonnegative Matrix Factorization), which performs cell type identification by combining sparse nonnegative matrix factorization (sparse NMF) with k-means clustering to cluster high-dimensional gene expression and low-dimensional location data. We show that, under multiple scenarios, including the cases where there is a small number of genes profiled and the location data is noisy, scHybridNMF outperforms sparse NMF, k-means, and an existing method that uses a hidden Markov random field to encode cell location and gene expression data for cell type identification.