Since its selection as the method of the year in 2013, single-cell technologies have become mature enough to provide answers to complex research questions. With the growth of single-cell profiling technologies, there has also been a significant increase in data collected from single-cell profilings, resulting in computational challenges to process these massive and complicated datasets. To address these challenges, deep learning (DL) is positioned as a competitive alternative for single-cell analyses besides the traditional machine learning approaches. Here, we survey a total of 25 DL algorithms and their applicability for a specific step in the single cell RNA-seq processing pipeline. Specifically, we establish a unified mathematical representation of variational autoencoder, autoencoder, generative adversarial network and supervised DL models, compare the training strategies and loss functions for these models, and relate the loss functions of these models to specific objectives of the data processing step. Such a presentation will allow readers to choose suitable algorithms for their particular objective at each step in the pipeline. We envision that this survey will serve as an important information portal for learning the application of DL for scRNA-seq analysis and inspire innovative uses of DL to address a broader range of new challenges in emerging multi-omics and spatial single-cell sequencing.

Deep learning has been applied in precision oncology to address a variety of gene expression-based phenotype predictions. However, gene expression data’s unique characteristics challenge the computer vision-inspired design of popular Deep Learning (DL) models such as Convolutional Neural Network (CNN) and ask for the need to develop interpretable DL models tailored for transcriptomics study. To address the current challenges in developing an interpretable DL model for modeling gene expression data, we propose a novel interpretable deep learning architecture called T-GEM, or Transformer for Gene Expression Modeling. We provided the detailed T-GEM model for modeling gene–gene interactions and demonstrated its utility for gene expression-based predictions of cancer-related phenotypes, including cancer type prediction and immune cell type classification. We carefully analyzed the learning mechanism of T-GEM and showed that the first layer has broader attention while higher layers focus more on phenotype-related genes. We also showed that T-GEM’s self-attention could capture important biological functions associated with the predicted phenotypes. We further devised a method to extract the regulatory network that T-GEM learns by exploiting the attributions of self-attention weights for classifications and showed that the network hub genes were likely markers for the predicted phenotypes.

There is a need to increase the performance and longevity of dental composites and accelerate the translation of novel composites to the market. This study explores the use of artificial intelligence (AI), specifically machine learning (ML) models, to predict the performance outcomes (POs) of dental composites from their composite attributes. A comprehensive dataset was carefully curated and refined from 200+ publications. Nine ML models were trained to predict discrete POs, and their performance was evaluated. Five models were used for regression analysis of continuous POs. Different ML models performed better on different POs. The k-nearest neighbors (KNN) model excelled in predicting flexural modulus (FlexMod), Decision Tree model in flexural strength (FlexStr) and volumetric shrinkage (ShrinkV), and Logistic Regression and Support Vector Machine models in shrinkage stress (ShrinkStr). Receiver-operating characteristic area under the curve analysis confirmed these results but found that Random Forest was more effective for FlexStr and ShrinkV, suggesting the possibility of Decision Tree overfitting the data. Regression analysis revealed that the voting regressor was superior for FlexMod and ShrinkV predictions, while Decision Tree Regression was optimal for FlexStr and ShrinkStr. Feature importance analysis indicated triethylene glycol dimethacrylate is a key contributor to FlexMod and ShrinkV, bisphenol A glycidyl dimethacrylate and urethane dimethacrylate to FlexStr, and depth of cure, degree of monomer-to-polymer conversion, and filler loading to ShrinkStr. There is a need to conduct a full analysis using multiple ML models because different models predict different POs better and for a large, comprehensive dataset to train robust AI models to facilitate the prediction and optimization of composite properties and support the development of new dental materials.

This study introduces TG-ME, an innovative computational framework that integrates transformer with graph variational autoencoder (GraphVAE) models for dissection of tumoral niches using spatial transcriptomics data and morphological images. TG-ME effectively identifies and characterizes niches in bench datasets and a high resolution NSCLC dataset. The pipeline consists in different stages that include normalization, spatial information integration, morphological feature extraction, gene expression quantification, single cell expression characterization, and tumor niche characterization. For this, TG-ME leverages advanced deep learning techniques that achieve robust clustering and profiling of niches across cancer stages. TG-ME can potentially provide insights into the spatial organization of tumor microenvironments (TME), highlighting specific niche compositions and their molecular changes along cancer progression. TG-ME is a promising tool for guiding personalized treatment strategies by uncovering microenvironmental signatures associated with disease prognosis and therapeutic outcomes.

Diabetic foot ulcers (DFUs) are a severe complication of diabetes mellitus (DM), which often lead to hospitalization and non-traumatic amputations in the United States. Diabetes prevalence estimates in South Texas exceed the national estimate and the number of diagnosed cases is higher among Hispanic adults compared to their non-Hispanic white counterparts. San Antonio, a predominantly Hispanic city, reports significantly higher annual rates of diabetic amputations compared to Texas. The late identification of severe foot ulcers minimizes the likelihood of reducing amputation risk. The aim of this study was to identify molecular factors related to the severity of DFUs by leveraging a multimodal approach. We first utilized electronic health records (EHRs) from two large demographic groups, encompassing thousands of patients, to identify blood tests such as cholesterol, blood sugar, and specific protein tests that are significantly associated with severe DFUs. Next, we translated the protein components from these blood tests into their ribonucleic acid (RNA) counterparts and analyzed them using public bulk and single-cell RNA sequencing datasets. Using these data, we applied a machine learning pipeline to uncover cell-type-specific and molecular factors associated with varying degrees of DFU severity. Our results showed that several blood test results, such as the Albumin/Creatinine Ratio (ACR) and cholesterol and coagulation tissue factor levels, correlated with DFU severity across key demographic groups. These tests exhibited varying degrees of significance based on demographic differences. Using bulk RNA-Sequenced (RNA-Seq) data, we found that apolipoprotein E (APOE) protein, a component of lipoproteins that are responsible for cholesterol transport and metabolism, is linked to DFU severity. Furthermore, the single-cell RNA-Seq (scRNA-seq) analysis revealed a cluster of cells identified as keratinocytes that showed overexpression of APOE in severe DFU cases. Overall, this study demonstrates how integrating extensive EHRs data with single-cell transcriptomics can refine the search for molecular markers and identify cell-type-specific and molecular factors associated with DFU severity while considering key demographic differences.