WYMM Tour: Seattle

Environmental monitoring (EM) of cleanroom facilities is a critical component of good manufacturing practice (GMP) compliant pharmaceutical manufacturing. To maintain effective sanitization, aseptic practice and EM programs, it is important to trend the microorganisms typical to the cleanroom from ambient environment, equipment, material and human foot-traffic through the cleanroom. At the Fred Hutch Cell Processing Facility (CPF), viable particulates recovered from EM samplings were historically mailed to a third-party vendor that performed MALDI-TOF mass spectrometry identification. Limitations of this process include long, often >30 days turn-around time (TAT) and limited species-level identification. These issues present difficulty in maintaining timely and accurate trends in environmental control. To eliminate these issues, the CPF Quality Control unit collaborated with the Genomics Shared Resource laboratory to develop and validate microbial identification using Oxford Nanopore sequencing. This approach used the Rapid Barcoding kit to prepare sequencing libraries which were then sequenced on a PromethION flow cell targeting >100K reads per library. Data were analyzed using the wf-bacterial-genomes and wf-metagenomes pipelines. After initial optimization studies of microbial DNA recovery from growth media plates and bottles, test sample DNA extracts of known microbial strains were submitted for testing (Genomics blinded to identity). The assay was successfully able to identify 3 unique bacteria and 2 fungi at the species level even at suboptimal DNA isolation yields with significantly short TAT. To determine the assay’s sensitivity, specificity and precision for polymicrobial colonization, DNA extracted from mixed colonies of up to 3 bacteria species were sequenced in single-blinded studies. All three bacteria were correctly distinguished by the assay. Future applications of this assay include integration into a USP <71> sterility testing workflow for identification of microbial contamination in contaminated raw materials or manufactured drug product.

Environmental monitoring (EM) of cleanroom facilities is a critical component of good manufacturing practice (GMP) compliant pharmaceutical manufacturing. To maintain effective sanitization, aseptic practice and EM programs, it is important to trend the microorganisms typical to the cleanroom from ambient environment, equipment, material and human foot-traffic through the cleanroom. At the Fred Hutch Cell Processing Facility (CPF), viable particulates recovered from EM samplings were historically mailed to a third-party vendor that performed MALDI-TOF mass spectrometry identification. Limitations of this process include long, often >30 days turn-around time (TAT) and limited species-level identification. These issues present difficulty in maintaining timely and accurate trends in environmental control. To eliminate these issues, the CPF Quality Control unit collaborated with the Genomics Shared Resource laboratory to develop and validate microbial identification using Oxford Nanopore sequencing. This approach used the Rapid Barcoding kit to prepare sequencing libraries which were then sequenced on a PromethION flow cell targeting >100K reads per library. Data were analyzed using the wf-bacterial-genomes and wf-metagenomes pipelines. After initial optimization studies of microbial DNA recovery from growth media plates and bottles, test sample DNA extracts of known microbial strains were submitted for testing (Genomics blinded to identity). The assay was successfully able to identify 3 unique bacteria and 2 fungi at the species level even at suboptimal DNA isolation yields with significantly short TAT. To determine the assay’s sensitivity, specificity and precision for polymicrobial colonization, DNA extracted from mixed colonies of up to 3 bacteria species were sequenced in single-blinded studies. All three bacteria were correctly distinguished by the assay. Future applications of this assay include integration into a USP <71> sterility testing workflow for identification of microbial contamination in contaminated raw materials or manufactured drug product.

Autosomal dominant polycystic kidney disease (ADPKD) is the most common heritable kidney with a large phenotypic variability that ranges from benign cysts to end-stage kidney disease. This phenotypic variability is best understood in the context of a genetic variant, but standard genetic testing fails to identify the molecular cause in 7% of individuals with a family history of ADPKD and up to 40% of those with a classic ADPKD phenotype but no family history. Clinical sequencing rely on short-read sequencing (SRS), which is limited in detecting variants in PKD1 due to its large size, high GC content, and the presence of multiple highly homologous pseudogenes. Furthermore, somatic mosaicism, a post-zygotic mutation present in only a subset of cells, may account for up to 10% of unresolved ADPKD cases. While genetic testing is typically performed on blood-derived DNA, variant detection in the affected tissue is critical for identifying mosaicism. Kidney biopsies carry risks and are rarely performed in ADPKD patients. Urine contains kidney epithelial cells which form cysts in ADPKD. This makes urine-derived DNA a valuable source for identifying disease-specific variants and epigenetic marks. Additionally, long-read sequencing (LRS) can overcome the limitations of SRS by reading DNA molecules up to hundreds of thousands of bases long, allowing it to capture structural variants and resolve variants in repetitive or low-complexity genomic regions that are inaccessible to SRS. LRS also provides phasing to determine which variants lie on the same chromosome which is critical for interpreting complex genotypes, and the ability to generate high-quality sequencing data from smaller amounts of DNA. We have developed a method to prepare high-quality DNA libraries from urine samples, successfully detecting and phasing a large deletion in the PKD1 gene. Applying LRS to urine samples promises to increase the diagnostic yield for ADPKD by identifying previously "hidden" genetic variants. Furthermore, we are investigating the role of DNA methylation patterns, correlating them with cyst severity in families with identical PKD1 variants but differing disease outcomes. This work offers a powerful, less-invasive approach to uncover the molecular basis of ADPKD, potentially identifying new therapeutic targets for this variable and challenging disease.

Extrachromosomal circular DNA (ecDNA), also called double minutes, is a special form of genetic alteration often seen in cancers. Its presence is strongly linked to aggressive growth, therapy resistance, and poor outcomes. EcDNA is made up of rearranged DNA from different chromosomes, creating highly complex structures that increase genetic diversity within tumors and drive their evolution. To study these circles, we used advanced long-read sequencing technologies, which can read very large and complicated DNA stretches with high accuracy. We also developed a new assembly strategy specifically designed to reconstruct complete ecDNA structures. Applying this approach to glioblastoma tumors, we deciphered ecDNA structures in detail, providing new insights into their organization and rearrangements. In this talk, I will highlight the power of long-read sequencing utilizing Oxford Nanopore Technologies platform to reveal the hidden complexity of ecDNA. By directly connecting ecDNA structure with cancer evolution and treatment resistance, our study expands understanding of tumor biology while pointing to potential vulnerabilities for future therapeutic strategies.

A significant portion of the biomolecular world remains ‘dark’, limiting our ability to fully understand and engineer living systems. Among the most complex and challenging components is the proteome — the vast collection of proteins in their varying modified states, which serve as the molecular machines of cells. In this talk, I will share progress from the Nivala lab on adapting the Oxford Nanopore Technologies platform for 'long-read' single-molecule protein sequencing. I will discuss our technique and how we can use it identify proteoforms with single-molecule resolution.

400M tons of plastic is manufactured annually, only 9% is recycled, and 360M tons of plastic becomes waste. The unsustainable use of plastics demands new technology solutions to increase circularity. Common plastics like polyethylene terephthalate (PET) face shortages of high-quality, affordable recycled resin, perpetuating fossil-fuel manufacture of virgin PET. Enzymatic PET recycling is a promising approach to address quality and cost limitations of legacy recycling methods and reduces greenhouse gas emissions of plastic production. However, engineering high-performance PET-degrading enzymes (PETases) is necessary for industrial scale enzymatic recycling. Artificial intelligence (AI)-enabled protein engineering has emerged as a transformational technology for enzyme optimization. Here we describe an AI-guided engineering platform to enhance plastic degrading enzymes. Our platform comprises novel microfluidics and custom Oxford Nanopore Technologies DNA sequencing pipelines. The platform enables screening of large enzyme libraries, generating extensive datasets. We leverage these data to train protein language models to guide enzyme optimization. Using our approach, we improved a PETase >450-fold in only 18 months. The optimized enzyme shows required techno-economic performance for commercialization, achieving >90% PET degradation at 50L scale in 24hrs. Our AI-guided platform will accelerate design of the next generation of enzymes for economical and sustainable recycling of PET and other plastics.