Russell S. Schwartz

Scaling and Generalization of Discrete Diffusion Models for Tumor Phylogenies

This study uses discrete diffusion models to generate tumor phylogenies, finding that mid-scale models outperform deeper ones and that diverse training improves cross-regime generalization.

Saturating hepatic clearance drives elevated cfDNA and fragment shortening in cancer

This study shows that elevated cfDNA and fragment shortening in cancer are due to saturated hepatic clearance, not tumor burden alone.

Theoretical estimates on the expected number of mutations for reconstructing clonal lineage trees

The paper estimates the number of mutations needed to reconstruct clonal lineage trees from large datasets, showing that modest mutation counts suffice for accurate reconstruction.

Computationally reconstructing the evolution of cancer progression risk

This study develops a computational method to infer cancer progression risk by analyzing tumor phylogenetics and mutation data from TCGA cohorts, revealing general mechanisms and cohort-specific variability.

Marker selection strategies for circulating tumor DNA guided by phylogenetic inference

The paper develops methods to improve marker selection for ctDNA analysis by refining phylogenetic models and quantifying clonal frequencies through optimization.

CITEgeist: Accurate deconvolution of spatial transcriptomics with same-slide proteomics reveals midkine as a secreted microenvironment modulator in ESR1 mutant breast cancer.

Deconvolution and Phylogeny Inference of Diverse Variant Types Integrating Bulk DNA-seq with Single-cell RNA-seq

This paper introduces TUSV-int, a method that integrates bulk DNA-seq and scRNA-seq to improve deconvolution and phylogenetic inference for cancer clonal structure.

Optimizing Design of Genomics Studies for Clonal Evolution Analysis

This paper develops a simulation-based optimization framework for designing genomic studies focused on clonal evolution, particularly in cancer, to optimize resource allocation and study design.

Interpretable deep learning for chromatin-informed inference of transcriptional programs driven by somatic alterations across cancers

The paper introduces CITRUS, a deep learning model that infers transcriptional programs influenced by somatic alterations across 17 cancer types using genomic, transcriptomic, and epigenomic data from The Cancer Genome Atlas.

A Simulator for Somatic Evolution Study Design

The paper introduces a simulator for generating synthetic data to optimize study designs for analyzing somatic evolution in cancer and other diseases.

ConTreeDP: A consensus method of tumor trees based on maximum directed partition support problem

ConTreeDP computes clonal consensus trees from cancer genomic data by solving the Maximum Directed Partition Support Consensus Tree problem, showing efficiency and accuracy.

De novo Prediction of Cell-Drug Sensitivities Using Deep Learning-based Graph Regularized Matrix Factorization

This paper introduces DeepGRMF, a deep learning model that predicts cell-drug sensitivities by integrating drug structures, signaling impacts, and cell states, outperforming existing methods on cancer datasets.

Tumor heterogeneity assessed by sequencing and fluorescence in situ hybridization (FISH) data

The paper develops computational methods combining sequencing and miFISH data to improve accurate inference of clonal copy number evolution in tumors.

Tumor Copy Number Deconvolution Integrating Bulk and Single-Cell Sequencing Data

The paper develops methods to deconvolve tumor copy number data by integrating bulk and single-cell sequencing data, improving accuracy over bulk data alone.