Science
Drives Us


At DZD, we are developing a sequencing-based rapid diagnostic that identifies, within hours, both the species and the antibiotic resistance profile of a bacterial pathogen. Current approaches take 2-5 days to provide similar information, a time delay that is associated with an 8% increase in death per hour for severe infections. Our diagnostic will enable hospitals to reduce patient length of stay, the overuse of expensive and often ineffective antibiotics, and most importantly, excess mortality.

Our diagnostic technologies are focused on three major areas that work in tandem with commercially available whole genome sequencing technology.

Sample Preparation

Sample preparation is a critical step to enabling pathogen genomic sequencing from clinical samples, and thereby avoiding the need for culture.

However, clinical samples, for example blood drawn from septic patients, present highly challenging material for downstream molecular processing that have made it difficult for technologies to advance in this space. Day Zero Diagnostics has developed a number of novel sample preparation technologies that open up the ability to sequence the genomes of pathogens directly from clinical samples.

Blood2Bac™

Blood2Bac™ is a process for ultra-high enrichment of bacterial DNA that converts a clinical blood sample with a billion times more human DNA than bacterial DNA to a diagnostic readout that contains far more bacterial DNA sequencing data than human data. Blood2Bac, which is patent protected, has been validated across over 50 species of bacterial pathogens to date at the clinically relevant concentration of 1 CFU/mL of blood. The Blood2Bac process enables high-depth, whole-genome sequence coverage of bacterial pathogens directly from clinical blood samples and is a critical technology for our sequencing-based diagnostic.

Computational Diagnostics

Computational Diagnostic Algorithms

While sample preparation technologies enable the whole genome sequencing of pathogens directly from clinical samples, once the samples have been sequenced there is a critical need for robust algorithms to interpret the sequencing data.

Our computational methods take the raw genomic data as input and transform them into the diagnostic results that can serve infectious disease clinicians and clinical microbiologists.

Keynome® ID

Keynome® ID is our proprietary algorithm that classifies the bacterial species present in a sample. Leveraging an internally-curated reference genome database to ensure high-accuracy across a large number pathogen species, Keynome ID provides high confidence for identifying the pathogen in clinical samples. When paired with the ultra-high enrichment of bacterial DNA using Blood2Bac preprocessing, Keynome ID determines the species of infections at concentrations as low as 1CFU/mL without being impaired by the high false-positive rates that often greatly reduce the diagnostic applicability and interpretability of hyper-sensitive molecular methods.

Keynome® gAST

Keynome® gAST is our machine learning algorithm for performing genomic Antibiotic Susceptibility Testing (gAST). Keynome gAST determines the antibiotic resistance and susceptibility profile of a pathogen directly from whole genome sequencing data. Keynome is a modern, novel machine-learning algorithm that leverages our proprietary training dataset and recent advances in the field of artificial intelligence in order to provide high accuracy gAST across a comprehensive range of species and antibiotic drugs. Unlike traditional “database look-up” approaches that identify the presence or absence of a limited set of experimentally validated resistance genes, Keynome allows for the simultaneous testing of a comprehensive set of bacterial species and resistance markers, along with the ability to identify novel resistance determinants as they emerge. Keynome gAST can be applied to the sequencing data of a clinical sample processed by Blood2Bac in order to determine the AST of a pathogen within hours of a blood draw.

Big Data

A prerequisite for the development of robust computational algorithms is access to a large, high-quality dataset. Day Zero Diagnostics has invested in the creation of such a dataset, an asset which uniquely enables the development and validation of highly-accurate genomic algorithms.

MicrohmDB®

MicrohmDB® is our large-scale, proprietary microbial database that combines the whole genome sequencing data of clinically relevant pathogens with their phenotypically derived resistance and susceptibility profiles. This dataset, one of the largest of its kind in the world, provides Day Zero with a unique ability to train our machine learning algorithms on an ever-growing base of clinically relevant data. The data in MicrohmDB undergoes rigorous curation and quality control checks, both for the sequencing and phenotypic data, thus ensuring that downstream computational applications have access to the highest quality data for training and validation.

Top