Cancer subtype classification has the potential to significantly improve disease prognosis and develop individualized patient management. Existing methods are limited by their ability to handle extremely high-dimensional data and by the influence of misleading, irrelevant factors, resulting in ambiguous and overlapping subtypes. To address the above issues, we proposed a novel approach to disentangling and eliminating irrelevant factors by leveraging the power of deep learning. Specifically, we designed a deep-learning framework, referred to as DeepType, that performs joint supervised classification, unsupervised clustering and dimensionality reduction to learn cancer-relevant data representation with cluster structure. We applied DeepType to the METABRIC breast cancer dataset and compared its performance to state-of-the-art methods. DeepType significantly outperformed the existing methods, identifying more robust subtypes while using fewer genes. The new approach provides a framework for the derivation of more accurate and robust molecular cancer subtypes by using increasingly complex, multi-source data. An open-source software package for the proposed method is freely available at http://www.acsu.buffalo.edu/~yijunsun/lab/DeepType.html. Supplementary data are available at Bioinformatics online.

Currently, taxonomic interrogation of microbiota is based on amplification of 16S rRNA gene sequences in clinical and scientific settings. Accurate evaluation of the microbiota depends heavily on the primers used, and genus/species resolution bias can arise with amplification of non-representative genomic regions. The latest Illumina MiSeq sequencing chemistry has extended the read length to 300 bp, enabling deep profiling of large number of samples in a single paired-end reaction at a fraction of the cost. An increasingly large number of researchers have adopted this technology for various microbiome studies targeting the 16S rRNA V3–V4 hypervariable region. To expand the applicability of this powerful platform for further descriptive and functional microbiome studies, we standardized and tested an efficient, reliable, and straightforward workflow for the amplification, library construction, and sequencing of the 16S V1–V3 hypervariable region using the new 2 × 300 MiSeq platform. Our analysis involved 11 subgingival plaque samples from diabetic and non-diabetic human subjects suffering from periodontitis. The efficiency and reliability of our experimental protocol was compared to 16S V3–V4 sequencing data from the same samples. Comparisons were based on measures of observed taxonomic richness and species evenness, along with Procrustes analyses using beta(β)-diversity distance metrics. As an experimental control, we also analyzed a total of eight technical replicates for the V1–V3 and V3–V4 regions from a synthetic community with known bacterial species operon counts. We show that our experimental protocol accurately measures true bacterial community composition. Procrustes analyses based on unweighted UniFrac β-diversity metrics depicted significant correlation between oral bacterial composition for the V1–V3 and V3–V4 regions. However, measures of phylotype richness were higher for the V1–V3 region, suggesting that V1–V3 offers a deeper assessment of population diversity and community ecology for the complex oral microbiota. This study provides researchers with valuable experimental evidence for the selection of appropriate 16S amplicons for future human oral microbiome studies. We expect that the tested 16S V1–V3 framework will be widely applicable to other types of microbiota, allowing robust, time-efficient, and inexpensive examination of thousands of samples for population, phylogenetic, and functional crossectional and longitutidal studies.

Conventional spectrometer designs necessitate a compromise between their resolution and sensitivity, especially as device and detector dimensions are scaled down. Here, we report on a miniaturizable spectrometer platform where light throughput onto the detector is instead enhanced as the resolution is increased. This planar, CMOS-compatible platform is based around metasurface encoders designed to exhibit photonic bound states in the continuum, where operational range can be altered or extended simply through adjusting geometric parameters. This system can enhance photon collection efficiency by up to two orders of magnitude versus conventional designs; we demonstrate this sensitivity advantage through ultralow-intensity fluorescent and astrophotonic spectroscopy. This work represents a step forward for the practical utility of spectrometers, affording a route to integrated, chip-based devices that maintain high resolution and SNR without requiring prohibitively long integration times.

Many scientific datasets are of high dimension, and the analysis usually requires retaining the most important structures of data. Principal curve is a widely used approach for this purpose. However, many existing methods work only for data with structures that are mathematically formulated by curves, which is quite restrictive for real applications. A few methods can overcome the above problem, but they either require complicated human-made rules for a specific task with lack of adaption flexibility to different tasks, or cannot obtain explicit structures of data. To address these issues, we develop a novel principal graph and structure learning framework that captures the local information of the underlying graph structure based on reversed graph embedding. As showcases, models that can learn a spanning tree or a weighted undirected `1 graph are proposed, and a new learning algorithm is developed that learns a set of principal points and a graph structure from data, simultaneously. The new algorithm is simple with guaranteed convergence. We then extend the proposed framework to deal with large-scale data. Experimental results on various synthetic and six real world datasets show that the proposed method compares favorably with baselines and can uncover the underlying structure correctly.

Sequence analysis is arguably a foundation of modern biology. Classic approaches to sequence analysis are based on sequence alignment, which is limited when dealing with large-scale sequence data. A dozen of alignment-free approaches have been developed to provide computationally efficient alternatives to alignment-based approaches. However, existing methods define sequence similarity based on various heuristics and can only provide rough approximations to alignment distances. In this article, we developed a new approach, referred to as SENSE (SiamEse Neural network for Sequence Embedding), for efficient and accurate alignment-free sequence comparison. The basic idea is to use a deep neural network to learn an explicit embedding function based on a small training dataset to project sequences into an embedding space so that the mean square error between alignment distances and pairwise distances defined in the embedding space is minimized. To the best of our knowledge, this is the first attempt to use deep learning for alignment-free sequence analysis. A large-scale experiment was performed that demonstrated that our method significantly outperformed the state-of-the-art alignment-free methods in terms of both efficiency and accuracy. Open-source software for the proposed method is developed and freely available at https://www.acsu.buffalo.edu/∼yijunsun/lab/SENSE.html. Supplementary data are available at Bioinformatics online.