David Koslicki

The paper introduces resources for large-scale metagenome comparison in the SRA, revealing trends in microbial discovery and improving assembly quality.

This paper introduces a semantic precision score to filter generic concepts from biomedical knowledge graphs, outperforming existing methods in drug repurposing queries.

This paper develops k-mer-based estimators for substitution, insertion, and deletion mutation rates, showing improved accuracy when accounting for indels in mutation models.

The paper introduces MetagenomicKG, a knowledge graph integrating metagenomic data with biomedical knowledge to enable hypothesis generation and pathogen prediction.

The paper benchmarks discrepancies across microbial genome reference databases, finding high consistency in viral genomes but significant variability in fungal genomes.

This paper introduces FracMinHash dN/dS, an alignment-free method to estimate dN/dS ratios across genomes, showing it is scalable and comparable to traditional methods.

The paper introduces a new sketching algorithm for genomic sequences that produces sub-linear sketches, balancing efficiency and accuracy between MinHash and FracMinHash.

The paper introduces a substring index called the Prokrustean graph to efficiently analyze k-mer-based quantities across different k-mer sizes.

Assembling the Sequence Read Archive (27.3 million accessions) with over 100-fold compression enables minutes-scale k-mer search and mining of enzymes, plasmids, proteins and viral reactivation.

The paper develops methods for assigning branch lengths to trees with fixed topology and applies them to improve functional analysis in metagenomics.

This paper introduces a theoretical framework for estimating cosine similarity using FracMinHash sketches and presents a fast implementation called frac-kmc.

YACHT provides a statistical test to determine microbial presence/absence in metagenomic samples, accounting for sequence divergence and incomplete sequencing depth.

The paper shows that open syncmer sketches are equivalent to FracMinHash sketches for k-mer similarities but offer better distance distribution and conservation, and extends them for metagenomic and sequence comparison applications.

This paper introduces fmh-funprofiler, a fast and lightweight metagenomic functional profiler using FracMinHash sketches, showing improved purity and speed compared to alignment-based methods.

This paper analyzes FracMinHash, showing it can be corrected for bias and used to estimate mutation distances with confidence intervals.

The paper introduces L2UniFrac, a modified UniFrac metric that allows for the computation of biologically meaningful averages of metagenomic samples.

ARAX is a graph-based reasoning tool that integrates knowledge across sources to answer translational biomedical questions.

KGML-xDTD predicts drug-disease treatments and explains mechanisms via knowledge graph paths, achieving state-of-the-art results in drug repurposing and MOA recapitulation.

RTX-KG2 integrates 70 biomedical knowledge sources into a standards-compliant knowledge graph with a Biolink schema, providing downloadable and queryable versions via API.

TAMPA provides an interpretable visualization tool for metagenomics taxonomic profiles, enabling better biological insights and tool selection.

The paper shows that the minimizer Jaccard estimator is biased and inconsistent, affecting the accuracy of sequence similarity estimation and a read mapping tool.

CMash provides a fast, multi-resolution method to estimate k-mer-based Jaccard and containment indices without reconstructing k-mer sets for different k values.

The paper evaluates metagenomic software tools using realistic datasets, highlighting improvements in assembly and taxon profiling while identifying remaining challenges in virus/archaea detection and reproducibility.

The paper analyzes the statistical properties of k-mers under a mutation process and develops hypothesis tests and confidence intervals for mutation rates using k-mer counts and Jaccard similarity.

This paper develops a method to compute the UniFrac metric on Whole Genome Shotgun metagenomic data by assigning branch lengths to taxonomic trees, showing it is robust and comparable to traditional 16S UniFrac.

Greedy minimum set cover over FracMinHash k-mer sketches selects a small set of reference genomes explaining a metagenome, supporting species-level taxonomic profiling and choice of genomes for read mapping.

This tutorial explains how to use the CAMI benchmarking toolkit to evaluate metagenomics software performance using standardized datasets and metrics.

Metalign improves metagenomic profiling by using containment min hash for efficient alignment-based analysis, achieving better results than existing methods on benchmark datasets.

The paper evaluates existing microbial databases and proposes guidelines for creating a unified reference database to enhance the quality and comprehensiveness of metagenomics research.

This paper introduces a method for metagenomic taxonomic profiling that minimizes biological diversity to improve reconstruction accuracy and computational efficiency.