We are thrilled to celebrate a major milestone for one of our core team members: Founding Scientist David Laub has successfully defended his PhD in Biomedical Informatics at UCSD under the mentorship of Hannah Carter!
David’s doctoral research addresses one of the fundamental barriers in computational genomics—the massive data storage and loading bottlenecks that hinder training deep learning models on personalized genomes. Through his PhD work and his key contributions at Standard Model Bio, David is fundamentally changing how biobank-scale genomic data is ingested and modeled. David lead key Computational Breakthroughs at UCSD: GenVarLoader & GenVarFormer.
During his PhD in Dr. Hannah Carter’s lab at UCSD, David developed novel architectures and dataloading frameworks designed to make biological sequence modeling efficient and accessible:
GenVarLoader: Published in BMC Bioinformatics, GenVarLoader is an accelerated dataloader that constructs personalized genomic sequences and functional tracks on-the-fly during training. Rather than writing uncompressed personalized genomes to disk, GenVarLoader reduces storage requirements approximately 2,000-fold—compressing the 1,000 Genomes Project dataset from 6.3 TB down to just 3.1 GB. It reads variants up to 26,000 times faster than pre-subset BCF files and loads data 130 times faster than conventional pipelines, keeping GPUs at near 100% utilization where traditional workflows bottleneck at ~30%.
GenVarFormer: To complement GenVarLoader, David designed GenVarFormer, a genomic transformer model that encodes only the specific variant sites where an individual differs from the reference genome. By scaling computational cost with genetic variation rather than window length, GenVarFormer drastically reduces the expense of training sequence models across entire populations.
Powering the Molecular Layer at Standard Model Bio
As Founding Scientist at Standard Model Bio, David leads a team developing breakthroughs, such as
— Whole-Genome Variant-to-Expression Modeling.
— Biobank-Scale Cloud Engineering.
— Open Science & Academic Collaborations
David’s PhD committee praised his defense for opening up dozens of exciting research avenues at the intersection of deep learning and human genetics. We could not be prouder to have Dr. David Laub continuing to drive molecular innovation at Standard Model Bio as we scale world-class predictive models for biomedicine.
David is an avid climber, yoga practitioner and bass guitarist. Please join us in congratulating Dr. David Laub!
Congrats!