Rachel Karchin

PICTographPlus infers clone-resolved transcriptional programs from bulk RNA-seq by integrating clonal phylogenies and proportions, achieving high accuracy in recovering clone-level expression and localizing pathway changes.

This study evaluates how different MHC representations affect neoantigen prediction performance, finding that curated pseudosequences are still most effective.

The paper identifies multi-omic phenotypes in pancreatic precursor lesions to distinguish progressive from indolent precancers using integrative transcriptomic, genomic, and spatial data.

The paper calibrates three new computational tools for clinical variant classification, finding they provide evidence strength comparable to previously recommended tools.

SVCFit estimates SV cellular fractions from bulk sequencing to reconstruct clonal evolution in mCRPC under therapy, showing treatment-driven clonal reconfiguration.

The paper assesses rare missense variants in genome sequences using PP3/BP4 evidence strength classifications and finds few variants meet strong or moderate criteria.

The paper assesses clustering of missense variants in neurodevelopmental disorders and cancer, identifying proteins with significant variation patterns in each condition.

The paper introduces BigMHC, a deep learning model that predicts MHC-I epitopes and immunogenic neoepitopes with high accuracy, outperforming existing methods.

The paper calibrates computational tools for missense variant pathogenicity classification and provides evidence-based revisions to PP3/BP4 ACMG/AMP criteria.

This paper introduces PICTograph, a Bayesian model for estimating cancer cell fractions and clone trees from multi-region sequencing data, showing improved accuracy in subclone reconstruction and evolutionary inference.

The paper analyzes HLA expression across 33 cancer types to show its role in predicting patient survival and immune evasion mechanisms.

OpenCRAVAT provides an open-source platform for variant and gene prioritization by integrating diverse data resources and computational methods.

CHASMplus identifies cancer-specific driver missense mutations across 32 cancer types using TCGA data, revealing over 4,000 unique mutations in 240 genes.

CRAVAT 4 provides tools for analyzing and interpreting somatic mutations in cancer studies, offering annotation, prioritization, and visualization with a web interface.

This study evaluates machine learning methods for predicting peptide-MHC Class I binding affinities, finding that MHCnuggets performs similarly to existing methods but is faster and handles variable-length peptides.

This paper introduces a framework for inferring subclonal hierarchies in cancer using somatic mutations, demonstrating improved power and error control in multi-region tumor sequencing.