Cerebral blood volume (CBV) is a hemodynamic correlate of oxygen metabolism and reflects brain activity and function. High-resolution CBV maps can be generated using the steady-state gadolinium-enhanced MRI technique. Such a technique requires an intravenous injection of exogenous gadolinium based contrast agent (GBCA) and recent studies suggest that the GBCA can accumulate in the brain after frequent use. We hypothesize that endogenous sources of contrast might exist within the most conventional and commonly acquired structural MRI, potentially obviating the need for exogenous contrast. Here, we test this hypothesis by developing and optimizing a deep learning algorithm, which we call DeepContrast, in mice. We find that DeepContrast performs equally well as exogenous GBCA in mapping CBV of the normal brain tissue and enhancing glioblastoma. Together, these studies validate our hypothesis that a deep learning approach can potentially replace the need for GBCAs in brain MRI.

While single-cell technologies provide snapshots of tumor states, building continuous trajectories and uncovering causative gene regulatory networks remains a significant challenge. We present Cflows , an AI framework that combines neural ODE networks with Granger causality to infer continuous cell state transitions and gene regulatory interactions from static scRNA-seq data. In a new 5-time point dataset capturing tumorsphere development over 30 days, Cflows reconstructs two types of trajectories leading to tumorsphere formation or apoptosis. Trajectory-based cell-of-origin analysis delineated a novel cancer stem cell profile characterized by CD44 ^hi EPCAM ⁺ CAV1 ⁺ , and uncovered a cell cycle–dependent enrichment of tumorsphere-initiating potential in G2/M or S-phase cells. Cflows uncovers ESRRA as a crucial causal driver of the tumor-forming gene regulatory network. Indeed, ESRRA inhibition significantly reduces tumor growth and metastasis in vivo. Cflows offers a powerful framework for uncovering cellular transitions and dynamic regulatory networks from static single-cell data.

Global warming, which largely results from excessive carbon emission, has become an increasingly heated international issue due to its ever-detereorating trend and the profound consequences. Plants sequester a large amount of atmospheric CO ${}_2$ via photosynthesis, thus greatly mediating global warming. In this study, we aim to model the temporal dynamics of photosynthesis for two different vegetation types to further understand the controlling factors of photosynthesis machinery. We experimented with a feedforward neural network that does not utilize past histories, as well as two networks that integrate past and present information, long short-term memory and transformer. Our results showed that one single climate driver, shortwave radiation, carries the most information with respect to prediction of upcoming photosynthetic activities. We also demonstrated that photosynthesis and its interactions with climate drivers, such as temperature, precipitation, radiation, and vapor pressure deficit, has an internal system memory of about two weeks. Thus, the predictive model could be best trained with historical data over the past two weeks and could best predict temporal evolution of photosynthesis two weeks into the future.

Alternative splicing profoundly diversifies the transcriptome and proteome, but decoding its regulatory mechanisms remains a challenge. We introduce CellSpliceNet, an interpretable transformer-based multimodal deep learning framework designed to predict splicing outcomes across the neurons of C. elegans. By integrating four complementary data modalities, namely long-range genomic sequence, local regions of interest (ROIs) in the RNA sequence, secondary structure, and gene expression, CellSpliceNet captures the complex interplay of factors that influence splicing decisions within the cellular context. CellSpliceNet employs modality-specific embeddings including different scales of RNA sequence (exon-flanking and whole-gene), generated RNA structure, and cell-type-specific gene expression. Gene expression and structure encoders employ expressive graph signal processing-based encoders that utilize the graph scattering transform. Further, CellSpliceNet introduces a novel multimodal multi-head attention mechanism that preserves the integrity of each modality while facilitating selective cross-modal interactions, notably allowing cellular gene expression to inform sequence and structural predictions. Attention-based pooling within each modality highlights biologically critical elements, such as canonical intron–exon splice boundaries and accessible single-stranded RNA loop structures within exons. We apply CellSpliceNet to a unique multimodal dataset measuring DNA sequencing, RNA sequencing, and alternative splicing frequencies in purified neuronal subtypes of the C. elegans. Our results show superior performance of CellSpliceNet compared to several other current models, and ablations demonstrate that each module of CellSpliceNet is essential for optimal performance.