Background The goal of the Fifth Annual Baylor College of Medicine & DNAnexus Structural Variation Hackathon was to push forward the research on structural variants (SVs) by rapidly developing and deploying open-source software. The event took place in-person and virtually in August 2023, when 49 scientists from 14 countries and 8 U.S. states collaboratively worked on projects to address critical gaps in the field of genomics. The hackathon projects concentrated on developing bioinformatic workflows for the following challenges: RNA transcriptome comparison, simulation of mosaic variations, metagenomics, Mendelian variation, SVs in plant genomics, and assembly vs. mapping SV calling comparisons. Methods As a starting point we used publicly available data from state-of-the-art long- and short-read sequencing technologies. The workflows developed during the hackathon incorporated open-source software, as well as scripts written using Bash and Python. Moreover, we leveraged the advantages of Docker and Snakemake for workflow automation. Results The results of the hackathon consists of six prototype bioinformatic workflows that use open-source software for SV research. We made the workflows scalable and modular for usability and reproducibility. Furthermore, we tested the workflows on example public data to show that the workflows can work. The code and the data produced during the event have been made publicly available on GitHub (https://github.com/collaborativebioinformatics) to reproduce and built upon in the future. Conclusions The following sections describe the motivation, lessons learned, and software produced by teams during the hackathon. Here, we describe in detail the objectives, value propositions, implementation, and use cases for our workflows. In summary, the article reports the advancements in the development of software for SV detection made during the hackathon.

In October 2020, 62 scientists from nine nations worked together remotely in the Second Baylor College of Medicine & DNAnexus hackathon, focusing on different related topics on Structural Variation, Pan-genomes, and SARS-CoV-2 related research. The overarching focus was to assess the current status of the field and identify the remaining challenges. Furthermore, how to combine the strengths of the different interests to drive research and method development forward. Over the four days, eight groups each designed and developed new open-source methods to improve the identification and analysis of variations among species, including humans and SARS-CoV-2. These included improvements in SV calling, genotyping, annotations and filtering. Together with advancements in benchmarking existing methods. Furthermore, groups focused on the diversity of SARS-CoV-2. Daily discussion summary and methods are available publicly at https://github.com/collaborativebioinformatics provides valuable insights for both participants and the research community.

In October 2020, 62 scientists from nine nations worked together remotely in the Second Baylor College of Medicine & DNAnexus hackathon, focusing on different related topics on Structural Variation, Pan-genomes, and SARS-CoV-2 related research.   The overarching focus was to assess the current status of the field and identify the remaining challenges. Furthermore, how to combine the strengths of the different interests to drive research and method development forward. Over the four days, eight groups each designed and developed new open-source methods to improve the identification and analysis of variations among species, including humans and SARS-CoV-2. These included improvements in SV calling, genotyping, annotations and filtering. Together with advancements in benchmarking existing methods. Furthermore, groups focused on the diversity of SARS-CoV-2. Daily discussion summary and methods are available publicly at  https://github.com/collaborativebioinformatics provides valuable insights for both participants and the research community.

The PacBio HiFi sequencing technology combines less accurate, multi-read passes from the same molecule (subreads) to yield consensus sequencing reads that are both long (averaging 10-25 kb) and highly accurate. However, these reads can retain residual sequencing error, predominantly insertions or deletions at homopolymeric regions. Here, we train deep learning models to polish HiFi reads by recognizing and correcting sequencing errors. We show that our models are effective at reducing these errors by 25-40% in HiFi reads from human as well as E. coli genomes.

Four years ago, an experimental system known as PilotNet became the first NVIDIA system to steer an autonomous car along a roadway. This system represents a departure from the classical approach for self-driving in which the process is manually decomposed into a series of modules, each performing a different task. In PilotNet, on the other hand, a single deep neural network (DNN) takes pixels as input and produces a desired vehicle trajectory as output; there are no distinct internal modules connected by human-designed interfaces. We believe that handcrafted interfaces ultimately limit performance by restricting information flow through the system and that a learned approach, in combination with other artificial intelligence systems that add redundancy, will lead to better overall performing systems. We continue to conduct research toward that goal. This document describes the PilotNet lane-keeping effort, carried out over the past five years by our NVIDIA PilotNet group in Holmdel, New Jersey. Here we present a snapshot of system status in mid-2020 and highlight some of the work done by the PilotNet group.

In October 2019, 46 scientists from around the world participated in the first National Center for Biotechnology Information (NCBI) Structural Variation (SV) Codeathon at Baylor College of Medicine. The charge of this first annual working session was to identify ongoing challenges around the topics of SV and graph genomes, and in response to design reliable methods to facilitate their study. Over three days, seven working groups each designed and developed new open-sourced methods to improve the bioinformatic analysis of genomic SVs represented in next-generation sequencing (NGS) data. The groups’ approaches addressed a wide range of problems in SV detection and analysis, including quality control (QC) assessments of metagenome assemblies and population-scale VCF files, de novo copy number variation (CNV) detection based on continuous long sequence reads, the representation of sequence variation using graph genomes, and the development of an SV annotation pipeline. A summary of the questions and developments that arose during the daily discussions between groups is outlined. The new methods are publicly available at https://github.com/NCBI-Codeathons/, and demonstrate that a codeathon devoted to SV analysis can produce valuable new insights both for participants and for the broader research community.