Alexander Dilthey

KIR*BLOOM improves KIR genotyping accuracy by modeling copy number and sequencing errors, achieving high precision and recall on whole-genome sequencing data.

The paper reprocesses SARS-CoV-2 genomic data to correct systematic errors in phylogenies using a novel assembly tool, resulting in a cleaner global tree of 4.47 million samples.

The paper uses Oxford Nanopore sequencing to identify genetic variation within the T. gondii ME49 strain, revealing copy number changes in the ROP8-ROP2A locus.

Hybrid HiFi/ultra-long ONT assemblies of 65 diverse humans yield 130 haplotypes, 1,246 complete centromeres, and a pangenome panel enabling short-read genome inference at median QV 45.

The paper introduces an integer programming framework for haplotype-resolved genotyping using pangenome graphs, achieving accurate MHC haplotype inference on low-coverage data.

This paper introduces Lemur and Magnet, lightweight tools for accurate taxonomic profiling of long-read metagenomic datasets, showing they run efficiently on laptops.

MetaGut uses long-read shotgun metagenomics to analyze gut microbiome composition and antibiotic resistance genes in patients undergoing alloHSCT, revealing significant inter-individual and temporal variability.

HOGVAX selects peptides to maximize population coverage in vaccine design by exploiting sequence overlaps, achieving high coverage and robust immunity against SARS-CoV-2 variants.

The paper completes six MHC haplotypes, including DR1, DR4, and structurally variable C4 regions, improving reference panels for immunogenomics analysis.

PanGenie improves variant calling by using a pangenome reference and k-mer counts, achieving higher accuracy and speed compared to traditional methods.

Haploflow assembles viral strain genomes from mixed samples, outperforming existing assemblers in speed and accuracy on benchmark datasets.

This paper introduces Swarm Learning, a decentralized machine learning approach for disease classification using transcriptomic data, demonstrating its effectiveness in developing classifiers for COVID-19, tuberculosis, and leukemias while preserving privacy.

Emu provides species-level microbial community profiling using full-length 16S rRNA reads with improved accuracy over existing methods.

The paper develops a diploid assembly-based benchmark for variants in the MHC region using HG002 data, identifying 21496 small variants with high concordance to clinical HLA types.

The paper analyzes HLA allele associations with 677 phenotypes in the UK Biobank, identifying 88 replicated and 90 novel associations, and uncovering non-additive and homozygous effects on disease risk.

This paper introduces Ultraplexing, a method that reduces sequencing and labor costs for hybrid bacterial genome assembly by 50% while maintaining high assembly quality.

This study identifies five genome-wide associations with severe malaria risk in 17,000 individuals from Africa, Asia, and Oceania, including a new variant linked to cerebral malaria.

MetaMaps assigns long-read metagenomic sequences to species and estimates sample composition with high accuracy and scalability.

The paper introduces an algorithm to validate distance constraints in graph-based genomic references using sparse matrix operations, demonstrating efficiency on real-world datasets.

The paper introduces a parallel algorithm for sequence-to-graph alignment, achieving significant speedup for aligning DNA reads to variation graphs.

HLA*PRG:LA is a graph alignment model for HLA typing that achieves high accuracy across various sequencing data types.

This paper introduces trio binning, a method for assembling haplotypes from heterozygous genomes using parental short reads, achieving high accuracy in diploid reconstruction.

The paper introduces a fast algorithm for computing whole-genome homology maps, achieving efficient alignment with high accuracy and coverage.

The paper introduces a fast approximate algorithm for mapping long reads to large reference databases, achieving high speed and precision by combining minimizers with MinHash identity estimation.

Nanopore-only sequencing of GM12878 across 39 MinION flowcells (91.2 Gb) yielded an assembly with NG50 ∼3 Mb; adding 5× ultra-long reads raised NG50 to 6.4 Mbp.

This paper develops a Bayesian framework to analyze genetic associations with healthcare phenotypes in the UK Biobank, identifying over 2,000 diagnostic term associations and novel immune-mediated disease links.

This paper introduces genome graphs as an alternative to linear reference genomes, showing they improve read mapping and variant calling compared to the standard GRCh38 reference.

The paper evaluates HLA imputation methods using a multi-ethnic panel, introduces a new method HLA*IMP:03, and shows it achieves high accuracy without requiring probability thresholds.

This paper develops HLA*PRG, a method for high-accuracy HLA type inference from whole-genome sequencing data, achieving accuracy comparable to gold-standard methods.

The paper introduces a population reference graph for improved genome inference in the MHC region, demonstrating enhanced accuracy using simulations and multiple data types.