Yaron Orenstein

This paper introduces an algorithm to compute minimum-density minimizers for high-throughput sequencing, achieving faster results than brute-force methods.

The paper benchmarks methods for inferring transcription factor binding specificities using a large community challenge and finds deep learning models outperform traditional methods, though positional weight matrices are competitive.

The paper constructs a comparative regulatory atlas of gonad development using mouse and human data, identifies a conserved enhancer critical for gonadal development, and prioritizes non-coding variants in DSD patients.

CROP predicts CRISPR-Cas9 frameshift rates using a context-aware, feature-independent method, outperforming existing approaches across multiple datasets.

This paper introduces 10-minimizers, a class of constant-space minimizers with lower density than random minimizers and efficient k-mer key retrieval.

The paper identifies context-dependent factors affecting CRISPR-Cas9 editing efficiency across species using endogenous editing data and shows that repair outcomes are conserved.

ExoShorkie improves RNA-seq coverage prediction for exogenous genomes in yeast using transfer learning, outperforming existing methods and revealing new regulatory insights.

GreedyMini generates DNA minimizers with low expected density, improving HTS analysis performance and approaching theoretical lower bounds.

The paper generates a comprehensive GUIDE-seq dataset with bulges, trains and evaluates machine-learning models to predict off-target sites with bulges, and visualizes model features.

The paper introduces new minimizer orders for large k-values using decycling sets, improving efficiency in DNA sequencing analysis.

The paper evaluates CRISPR off-target site prediction methods using large-scale in vitro and in vivo datasets, highlighting the importance of data preprocessing, inactive site inclusion, and mismatch features.

The paper demonstrates that using universal hitting sets improves de Bruijn graph construction efficiency by reducing runtime and memory usage in genome assembly.

The paper uses high-throughput oligo libraries and machine learning to design non-repetitive RNA binding cassettes for gene editing and RNA tracking, demonstrating their functionality in mammalian cells.

DeCoDe optimizes degenerate codon libraries for protein-coding DNA, outperforming existing methods and supporting mixed-length libraries.

This paper introduces PASHA, a randomized parallel algorithm for generating near-optimal universal hitting sets, which is significantly faster and more memory-efficient than existing methods.

This paper introduces DLPRB, a deep learning model for predicting protein-RNA interactions by analyzing RNA sequence and structure, showing improved accuracy over existing methods.

The paper improves minimizers and winnowing schemes by analyzing k-mer ordering and introducing universal hitting sets to enhance performance.