Jian Ma

EMBER: Contextualized gene representations enable compositional prediction across cellular perturbations

EMBER, a gene-centric foundation model, improves cross-context perturbation-response prediction and genetic-interaction ranking using contextualized gene embeddings.

Germline hypomethylation shapes dynamic CpG reservoirs in ape genomes

The paper shows that peri/centromeric regions in ape genomes are hypomethylated CpG reservoirs, linking germline hypomethylation to CpG evolution and structural transcriptomic innovation.

A meta-interaction basis for cell-cell communication in tissues

SpiderNet identifies cell-cell communication programs from spatial transcriptomics data, revealing key interactions and their associations with phenotypes and therapeutic responses.

DANGO: Predicting higher-order genetic interactions

This paper introduces a computational method to predict higher-order genetic interactions, demonstrating its effectiveness in yeast by predicting trigenic interactions and uncovering biological functions related to cell growth.

Uncertainty-aware synthetic lethality prediction with pretrained foundation models

This paper predicts synthetic lethal gene pairs using a foundation model-based framework with uncertainty calibration, improving performance without relying on PPI or GO data.

Dose-dependent NFI regulation of progenitor lifespan and output underlying human neocortical malformation

This study shows that NFI proteins regulate radial glial progenitor lifespan and output in a dose-dependent manner, affecting neocortical development and causing malformations in human cerebral organoids and mouse models.

An integrated view of the structure and function of the human 4D nucleome

The paper integrates 4D Nucleome data to analyze 3D genome structure and function in human cells, revealing chromatin domains, loops, and interactions, and showing how genetic variants affect genome folding.

Unified integration of spatial transcriptomics across platforms

LO_SCPLOWLOKIC_SCPLOW integrates spatial transcriptomics data across platforms without shared gene panels, improving cross-technology spatial gene program identification and tissue alignment.

MIMYR: Generative modeling of missing tissue in spatial transcriptomics

MIMYR reconstructs missing tissue regions in spatial transcriptomics data using a generative framework that predicts cell locations, assigns cell types, and generates gene expression profiles.

TissueNarrator: Generative Modeling of Spatial Transcriptomics with Large Language Models

This paper introduces a framework that uses large language models to generate and analyze spatial transcriptomics data, improving the simulation of cell behavior and prediction of interactions.

DNALONGBENCH: A Benchmark Suite for Long-Range DNA Prediction Tasks

The paper introduces DNALO_SCPLOWONGC_SCPLOWBO_SCPLOWENCHC_SCPLOW, a benchmark suite for long-range DNA prediction tasks, and evaluates five methods on five genomics tasks.

HEIMDALL: Disentangling tokenizer design for robust transfer in single-cell foundation models

This paper introduces HO_SCPLOWEIMDALLC_SCPLOW, a framework to analyze tokenizer design in scFMs, showing that specific tokenization choices improve generalization under distribution shift.

DANCE 2.0: Transforming single-cell analysis from black box to transparent workflow

DANCE 2.0 automates and interprets single-cell preprocessing by recommending optimal pipelines for various tasks, improving performance across six major analyses.

EYKTHYR reveals transcriptional regulators of spatial gene programs

EO_SCPLOWYKTHYRC_SCPLOW identifies spatial transcriptional regulators by integrating gene expression and chromatin accessibility in spatial multiome data, outperforming existing methods in mouse brain and tumor datasets.

POPARI: Modeling multisample variation in spatial transcriptomics

This paper introduces a probabilistic model for analyzing multisample spatial transcriptomics data, identifying spatial metagenes and interactions across tissues.

Beyond A and B Compartments: how major nuclear locales define nuclear genome organization and function

The paper examines genome organization in human cells relative to nuclear locales, finding that gene expression and replication timing correlate with different nuclear structures and that LADs and iLADs compete for lamina attachment.

Complete sequencing of ape genomes

The paper presents haplotype-resolved genomes of six ape species with high accuracy and provides comparative analyses to study evolutionary diversity and gene family evolution.

Frequent, context-dependent effects of human genetic variation on Cas9 activity revealed by population-scale GUIDE-seq-2 and deep combinatorial CHANCE-seq profiling

The paper uses population-scale GUIDE-seq-2 and CHANCE-seq to analyze how genetic variants affect Cas9 activity, finding frequent context-dependent effects on off-target sites.

L2G: Repurposing Language Models for Genomics Tasks

This paper introduces L2G, a method that repurposes pre-trained language models for genomics tasks, achieving better performance than specialized models on multiple benchmarks.

A Silicon Diode based Optoelectronic Interface for Bidirectional Neural Modulation

This paper demonstrates a silicon diode device that bidirectionally modulates neural activity in vitro and in vivo through photothermal and photovoltaic effects.

scGHOST: Identifying single-cell 3D genome subcompartments

SCGHOST identifies single-cell 3D genome subcompartments using scHi-C data and provides insights into cell-to-cell variability in nuclear organization.

Genome-wide analysis of the interplay between chromatin-associated RNA and 3D genome organization in human cells

This study investigates how chromatin-associated RNA interacts with 3D genome organization in human cells, revealing a feedback loop between RNA distribution and chromatin structure.

PhyGCN: Pre-trained Hypergraph Convolutional Neural Networks with Self-supervised Learning

PhyGCN improves node representations in hypergraphs using self-supervised learning and variable hyperedge sizes, showing effectiveness on chromatin interactions and polypharmacy side-effects.

Concurrent profiling of multiscale 3D genome organization and gene expression in single mammalian cells

This paper introduces GAGE-seq, a single-cell method that simultaneously profiles 3D genome structure and gene expression, revealing their relationship in defining cell identity and lineage commitment.

Best Practices for Interpretable Machine Learning in Computational Biology

This paper introduces a workflow for best practices in interpretable machine learning to enhance knowledge discovery in computational biology.

Ultrafast and interpretable single-cell 3D genome analysis with Fast-Higashi

Fast-Higashi improves scHi-C analysis by jointly identifying cell identities and chromatin meta-interactions with tensor decomposition and a novel algorithm, enhancing rare cell type detection and developmental trajectory reconstruction.

Nucleome Browser: An integrative and multimodal data navigation platform for 4D Nucleome

Nucleome Browser integrates heterogeneous 4D Nucleome data into an interactive platform for visualization and exploration across biological contexts.

CONCERT: Genome-wide prediction of sequence elements that modulate DNA replication timing

The paper introduces a machine learning method to predict DNA replication timing from sequence features and identifies genome-wide sequence elements that modulate replication timing.

Multiscale and integrative single-cell Hi-C analysis with Higashi

Higashi improves scHi-C data analysis by enhancing quality and extracting 3D genome features, revealing cell-to-cell variability and new gene regulation connections in human prefrontal cortex cells.

Distinct progenitor behavior underlying neocortical gliogenesis related to tumorigenesis

The paper quantifies the progenitor cell behavior and gliogenesis program in the mouse neocortex, revealing a 60%:15%:25% propensity for astrocytes, oligodendrocytes, and both, and links gliogenesis to tumor suppressor loss.