Search for: All records

We created GNQA, a generative pre-trained transformer (GPT) knowledge base driven by a performant retrieval augmented generation (RAG) with a focus on aging, dementia, Alzheimer’s and diabetes. We uploaded a corpus of three thousand peer reviewed publications on these topics into the RAG. To address concerns about inaccurate responses and GPT ‘hallucinations’, we implemented a context provenance tracking mechanism that enables researchers to validate responses against the original material and to get references to the original papers. To assess the effectiveness of contextual information we collected evaluations and feedback from both domain expert users and ‘citizen scientists’ on the relevance of GPT responses. A key innovation of our study is automated evaluation by way of a RAG assessment system (RAGAS). RAGAS combines human expert assessment with AI-driven evaluation to measure the effectiveness of RAG systems. When evaluating the responses to their questions, human respondents give a “thumbs-up” 76% of the time. Meanwhile, RAGAS scores 90% on answer relevance on questions posed by experts. And when GPT-generates questions, RAGAS scores 74% on answer relevance. With RAGAS we created a benchmark that can be used to continuously assess the performance of our knowledge base. Full GNQA functionality is embedded in the freeGeneNetwork.orgweb service, an open-source system containing over 25 years of experimental data on model organisms and human. The code developed for this study is published under a free and open-source software license athttps://git.genenetwork.org/gn-ai/tree/README.md. 

Pangenome graphs can represent all variation between multiple reference genomes, but current approaches to build them exclude complex sequences or are based upon a single reference. In response, we developed the PanGenome Graph Builder, a pipeline for constructing pangenome graphs without bias or exclusion. The PanGenome Graph Builder uses all-to-all alignments to build a variation graph in which we can identify variation, measure conservation, detect recombination events and infer phylogenetic relationships. 

Abstract Pangenome graphs can represent all variation between multiple genomes, but existing methods for constructing them are biased due to reference-guided approaches. In response, we have developed PanGenome Graph Builder (PGGB), a reference-free pipeline for constructing unbi-ased pangenome graphs. PGGB uses all-to-all whole-genome alignments and learned graph embeddings to build and iteratively refine a model in which we can identify variation, measure conservation, detect recombination events, and infer phylogenetic relationships. 

Abstract The short arms of the human acrocentric chromosomes 13, 14, 15, 21 and 22 (SAACs) share large homologous regions, including ribosomal DNA repeats and extended segmental duplications 1,2 . Although the resolution of these regions in the first complete assembly of a human genome—the Telomere-to-Telomere Consortium’s CHM13 assembly (T2T-CHM13)—provided a model of their homology 3 , it remained unclear whether these patterns were ancestral or maintained by ongoing recombination exchange. Here we show that acrocentric chromosomes contain pseudo-homologous regions (PHRs) indicative of recombination between non-homologous sequences. Utilizing an all-to-all comparison of the human pangenome from the Human Pangenome Reference Consortium 4 (HPRC), we find that contigs from all of the SAACs form a community. A variation graph 5 constructed from centromere-spanning acrocentric contigs indicates the presence of regions in which most contigs appear nearly identical between heterologous acrocentric chromosomes in T2T-CHM13. Except on chromosome 15, we observe faster decay of linkage disequilibrium in the pseudo-homologous regions than in the corresponding short and long arms, indicating higher rates of recombination 6,7 . The pseudo-homologous regions include sequences that have previously been shown to lie at the breakpoint of Robertsonian translocations 8 , and their arrangement is compatible with crossover in inverted duplications on chromosomes 13, 14 and 21. The ubiquity of signals of recombination between heterologous acrocentric chromosomes seen in the HPRC draft pangenome suggests that these shared sequences form the basis for recurrent Robertsonian translocations, providing sequence and population-based confirmation of hypotheses first developed from cytogenetic studies 50 years ago 9 . 

Abstract Here the Human Pangenome Reference Consortium presents a first draft of the human pangenome reference. The pangenome contains 47 phased, diploid assemblies from a cohort of genetically diverse individuals 1 . These assemblies cover more than 99% of the expected sequence in each genome and are more than 99% accurate at the structural and base pair levels. Based on alignments of the assemblies, we generate a draft pangenome that captures known variants and haplotypes and reveals new alleles at structurally complex loci. We also add 119 million base pairs of euchromatic polymorphic sequences and 1,115 gene duplications relative to the existing reference GRCh38. Roughly 90 million of the additional base pairs are derived from structural variation. Using our draft pangenome to analyse short-read data reduced small variant discovery errors by 34% and increased the number of structural variants detected per haplotype by 104% compared with GRCh38-based workflows, which enabled the typing of the vast majority of structural variant alleles per sample.