993. A CRISPR-Powered Universal Infectious Disease Assay

Abstract Background The COVID-19 pandemic has brought awareness to the dangers of emerging pathogens to global human health and welfare. Sensitivity and flexibility are important features for methods used to detect emerging pathogens. While PCR testing is rapid and sensitive, a significant advantage next generation sequencing (NGS) approaches have over PCR-based analyses is the ability to detect previously undiscovered pathogens while also providing genomic information that can detect SARS-CoV-2 genome sequence, identify source of co-infection, and the host transcriptional response in a single workflow. The critical component enabling this approach is Jumpcode CRISPRclean technology which removes abundant human and bacterial ribosomal RNA sequences from NGS libraries. CRISPRclean workflow easily integrates into next generation sequencing projects Schematic of the Jumpcode CRISPRclean protocol Methods CRISPRclean was applied to contrived infected tissue samples including human lung RNA spiked with serially diluted amounts of SARS-CoV-2 RNA and bacterial RNA. NEB RNA libraries were prepared and treated with CRISPRclean protocol, then sequenced on Illumina instruments. Data analysis was performed using Jumpcode proprietary software to measure alignment and depletion rates, the Silva database for rRNA read alignment, and Kraken2 and CosmosID pipelines for k-mer based metagenomic investigation. Fold enrichment of SARS-CoV-2 reads after CRISPRclean depletion of libraries prepared from contrived samples. CRISPRclean treatment of the fully contrived samples increases the fraction of reads that map to the SARS-CoV-2 genome by an average of ~10-fold Results CRISPRclean treatment of the contrived samples increases ~10 fold of reads that map to the SARS-CoV-2 genome. For the 60 viral copies of SARS-CoV-2 sample, the number of reads mapping to the SARSCoV-2 genome increases from ~10,000 reads to ~70,000 reads. A similar increase in reads occurs for S. aureus. The percentage of SARS-CoV-2 genome covered at 1X and 10X also increases. Similar results were achieved even after downsampling the datasets to 5M reads. There is a ~4-fold increase in bacterial species detection in these stool samples after CRISPRclean treatment. Percentage of SARS-CoV-2 genome covered at 1X and 10X increases as a result of rRNA depletion. Coverage of the SARS-CoV-2 genome at 50 million reads. Number of reads aligning to the S. aureus and SARS-CoV-2 genomes increases after CRISPRclean depletion. For the sample containing 0.0001% SARS-CoV-2, (60 viral copies), the number of reads mapping to the SARS-CoV-2 genome increases from ~10,000 reads to ~70,000 reads. CosmosID Shotgun Metagenomics Analysis heat map showing read alignments to viral genomes. The yellow color indicates high read counts. The CosmosID shotgun metagenomic analysis software was used to analyze the sequencing data, classify the sequences and generate the heat map. Conclusion Metatranscriptomics powered by CRISPR-mediated rRNA depletion offers a robust methodology to acquire viral genomic data, microbiome composition, co-infection information, and the transcriptional status of the host immune response in a single workflow. This sequencing-based approach can be available on the first day of the next viral outbreak and should be considered as a first-line test for novel zoonotic virus detection. Bacterial species composition of patient stool samples before and after CRISPRclean depletion. ~4-fold increase in bacterial species detection in these stool samples after CRISPRclean treatment. Sequencing data downsampled to 20 million reads. Disclosures Keith Brown, n/a, Jumpcode Genomics (Board Member, Employee, Shareholder)

Background. The COVID-19 pandemic has brought awareness to the dangers of emerging pathogens to global human health and welfare. Sensitivity and flexibility are important features for methods used to detect emerging pathogens. While PCR testing is rapid and sensitive, a significant advantage next generation sequencing (NGS) approaches have over PCR-based analyses is the ability to detect previously undiscovered pathogens while also providing genomic information that can detect SARS-CoV-2 genome sequence, identify source of co-infection, and the host transcriptional response in a single workflow. The critical component enabling this approach is Jumpcode CRISPRclean technology which removes abundant human and bacterial ribosomal RNA sequences from NGS libraries.
CRISPRclean workflow easily integrates into next generation sequencing projects Schematic of the Jumpcode CRISPRclean protocol Methods. CRISPRclean was applied to contrived infected tissue samples including human lung RNA spiked with serially diluted amounts of SARS-CoV-2 RNA and bacterial RNA. NEB RNA libraries were prepared and treated with CRISPRclean protocol, then sequenced on Illumina instruments. Data analysis was performed using Jumpcode proprietary software to measure alignment and depletion rates, the Silva database for rRNA read alignment, and Kraken2 and CosmosID pipelines for k-mer based metagenomic investigation. Fold enrichment of SARS-CoV-2 reads after CRISPRclean depletion of libraries prepared from contrived samples.
CRISPRclean treatment of the fully contrived samples increases the fraction of reads that map to the SARS-CoV-2 genome by an average of ~10-fold Results. CRISPRclean treatment of the contrived samples increases ~10 fold of reads that map to the SARS-CoV-2 genome. For the 60 viral copies of SARS-CoV-2 sample, the number of reads mapping to the SARSCoV-2 genome increases from ~10,000 reads to ~70,000 reads. A similar increase in reads occurs for S. aureus. The percentage of SARS-CoV-2 genome covered at 1X and 10X also increases. Similar results were achieved even after downsampling the datasets to 5M reads. There is a ~4-fold increase in bacterial species detection in these stool samples after CRISPRclean treatment. Percentage of SARS-CoV-2 genome covered at 1X and 10X increases as a result of rRNA depletion.
Coverage of the SARS-CoV-2 genome at 50 million reads. Number of reads aligning to the S. aureus and SARS-CoV-2 genomes increases after CRISPRclean depletion.
For the sample containing 0.0001% SARS-CoV-2, (60 viral copies), the number of reads mapping to the SARS-CoV-2 genome increases from ~10,000 reads to ~70,000 reads.
CosmosID Shotgun Metagenomics Analysis heat map showing read alignments to viral genomes. The yellow color indicates high read counts. The CosmosID shotgun metagenomic analysis software was used to analyze the sequencing data, classify the sequences and generate the heat map.
Conclusion. Metatranscriptomics powered by CRISPR-mediated rRNA depletion offers a robust methodology to acquire viral genomic data, microbiome composition, co-infection information, and the transcriptional status of the host immune response in a single workflow. This sequencing-based approach can be available on the first day of the next viral outbreak and should be considered as a first-line test for novel zoonotic virus detection. Bacterial species composition of patient stool samples before and after CRISPRclean depletion. 4-fold increase in bacterial species detection in these stool samples after CRISPRclean treatment. Sequencing data downsampled to 20 million reads.
Disclosures. Keith Brown, n/a, Jumpcode Genomics (Board Member, Employee, Shareholder) Background. Procalcitonin (PCT) and serum lactate (L) are measures of bacterial infection and tissue hypoxia, respectively, but also used to discern sepsis from infection negative systemic inflammation (INSI). However, improved tools are needed to enhance this differentiation. A previously validated gene signature assay (SeptiCyte RAPID) and its correlated score (SeptiScore (SS)) has been reported to effectively differentiate sepsis from INSI.

Comparison of Lactate, Procalcitonin and a Gene Signature Assay Alone or in Combination to Differentiate Sepsis from Non-infectious Systemic Inflammation in ICU Patients
Objective. To compare early L, PCT and SS results (alone or in combination) in differentiating sepsis from INSI in adult intensive care unit (ICU) patients (Pt).

L, PCT, SS Comparison of Sepsis vs INSI
Conclusion. L is sub-optimal in discriminating sepsis from INSI. PCT with or without L was acceptable but not as robust as SS. SS alone or in any combination provided superior and significant discrimination between sepsis and INSI. Incorporation of SS into the clinical assessment process for suspected sepsis pts should be evaluated to determine the impact on early detection and Pt management.

A Murine Model of Klebsiella pneumoniae Gastrointestinal Colonization with Parenteral Vancomycin Administration
Bettina Cheung, BS 1 ; Marine Lebrun-Corbin, BS, MS 1 ; Alan R. Hauser, MD PhD 1 ; 1 Northwestern University, Chicago, Illinois Session: P-56. Microbial Pathogenesis Background. As a leading cause of nosocomial infections, Klebsiella pneumoniae poses a significant threat due to its propensity to acquire resistance to many classes of antibiotics, including carbapenems. Gastrointestinal (GI) colonization by K. pneumoniae is a risk factor for subsequent infection as well as transmission to other patients. To study this crucial step in pathogenesis, we developed a mouse model of K. pneumoniae GI colonization using a clinically relevant parenteral antibiotic regimen.
Methods. To improve the clinical relevance of our model, we elected to use intraperitoneal injections of vancomycin, one of the most highly utilized antibiotics in the United States.
Results. To optimize dosage in C57bl/6 mice, we injected 20mg/kg, 350mg/kg, or vehicle (PBS) for three days prior to gastric gavage with 10 8 colony forming units (CFU) of a low-resistance strain of K. pneumoniae. The mice who received 350mg/kg (a mouse equivalent of a human dose of 1g/day calculated through the FDA guidelines for estimating safe dosing) shed about 10 7 CFU/g of feces at Day 7 while those receiving the lower dose or vehicle shed 10 4 CFU/g. Next, we compared 3-or 5-day pre-treatments with vancomycin prior to inoculation with an ST258 (epidemic carbapenem-resistant) strain. At Day 7 post-inoculation, mice who received 5 days shed 10 10 CFU/g feces while those who received vancomycin for 3 days or vehicle for 5 days (PBS) shed 10 6 or 10 4 CFU/g feces respectively. Thus, we chose 5 days of 350mg/kg vancomycin injection as our regimen for inducing robust GI colonization in mice. Finally, we tested the durability of colonization by following fecal shedding in mice up to Day 60 post-inoculation with a second ST258 strain. Shedding during the first 7 days occurs at about 10 10 CFU/g feces, and from day 14 to day 60 fecal loads are stable around 10 7 CFU/g feces. Results are comparable between male and female mice.
Conclusion. In conclusion, we have developed a mouse model of robust, prolonged GI colonization with multiple strains of K. pneumoniae using controlled dosing of a clinically relevant antibiotic. This model may be used to study a key step in K. pneumoniae pathogenesis and infection prevention in the future.
Disclosures. All Authors: No reported disclosures