#9 · ML / AI methods
Organ- and region-resolved models extract aging, disease, and genetic signals from omics…
The dominant thread is stratifying biological signal by organ, anatomical region, or cell context, then using learned models to relate it to age, disease state, or genotype. Organ-specific proteomic clocks for eight mouse organs are applied to RNA and spatial transcriptomes, with organ-dependent age signals. ATLAS maps 40M cells across mouse brains to compare cell-type distributions by genetic background. A head MRI network estimates calvarial bone marrow adiposity and links it to heritability and 41 loci. In Alzheimer's disease, ADVISE and ADInteractome use interpretable models on snRNA-seq to recover Braak-stage signals and gene interactions, with ROBO2 validated by CRISPRa. The ALS proteomics batch correction and NEPTUNE are adjacent data-quality and pathway-inference tools. The newer element is coupling resolved readouts to genetic or mechanistic validation.
Published papers
Spatial single-cell mapping of transcriptional differences across genetic backgrounds in mouse brains.
ATLAS, a scalable spatial mapping method using scRNA-seq-derived probe signatures, profiles 40M cells across mouse brains to compare cell-type distributions by…
ADInteractome: discovering Alzheimer's disease-associated gene interactions from single-nucleus RNA-seq data through neural interaction states.
ADInteractome is an interpretable AI framework that extracts Alzheimer's-associated gene interactions from neural network internal states trained on snRNA-seq…
Estimating bone marrow adiposity from head MRI and identifying its genetic architecture.
Neural network localizes calvarial bone marrow in head MRI to estimate adiposity, revealing heritability, 41 genetic loci, and genetic correlations with traits.
Preprints
Proteomic clocks reveal organ-specific signatures of aging in mice
Organ-specific proteomic aging clocks for eight mouse organs, applied to RNA and spatial transcriptomes with organ-dependent age signals.
Batch Effect Correction to Enhance Utility of a Publicly Available ALS Proteomics Dataset
Identifies and corrects a batch effect in the Answer ALS proteomic dataset (508 samples, 3,643 proteins) using BDC controls with ComBat, and releases a…
NEPTUNE: a neuronal cell death pathway prediction tool used to decipher neurodegeneration using single cell RNA-seq datasets
NEPTUNE is a user-friendly tool that identifies neuronal cell death markers from scRNA-seq differentially expressed gene lists, with pathway mapping,…
Disentangling Context-Specific Alzheimer's Disease Progression Signatures through an Explainable AI Framework
Interpretable deep generative model ADVISE disentangles Braak-stage signals in AD snRNA-seq; prioritizes ROBO2, validated by CRISPRa.
Topic history
| Date | Rank |
|---|---|
| #9 | |
| #38 | |
| #27 | |
| #7 | |
| #3 | |
| #9 | |
| #24 | |
| #12 | |
| #19 | |
| #12 | |
| #2 | |
| #20 | |
| #14 | |
| #19 | |
| #12 | |
| #2 | |
| #28 | |
| #36 | |
| #35 | |
| #33 | |
| #38 |