Lab Matters Winter 2021 | Page 24


Emerging Cloud-based Solutions for Newborn Screening Workflows

By Bryce Asay , PhD , newborn screening bioinformatics and data analytics fellow , Utah Public Health Laboratories ; Nicole Ruiz-Schultz , PhD , bioinformatician , Utah Public Health Laboratories and Andreas Rohrwasser , PhD , MBA , director , Newborn Screening , Utah Public Health Laboratories
Newborn screening ( NBS ) is an essential public health service that has successfully detected , diagnosed , and prevented serious and life-threatening outcomes in babies with genetic and metabolic diseases . On average , nearly 4 million babies born in the United States benefit from screening for up to 60 disorders every year . Screening is performed primarily using biochemical assays , with second-tier and confirmatory testing increasingly utilizing nextgeneration ( NGS ) or Sanger sequencing and genomic variant analysis ( GVA ). In biochemical assays , the concentration of a biochemical marker is measured and the value generated is compared relative to populations with and without the disease . With GVA , whole genomes or whole exomes are sequenced and variants of interest are identified using various databases containing variants that have been correlated with the disorder of interest . These approaches are computationally extensive , expensive and often require experts trained for a particular technology .
A New Future
When current technologies were originally adopted , NBS laboratories relied on on-site computing infrastructure , laboratory information management systems ( LIMS ) and network-attached storage devices that were maintained and run by information technology specialists , health informaticists and bioinformaticians . The system was optimized to handle peak load times to prevent bottlenecking . As a result , knowledge and resources became compartmentalized to specific individuals and departments . Today , much of the initial program-specific infrastructure is aging and the NBS workforce has failed to retain and / or recruit individuals with current training and knowledge of technologies and algorithms . Consequently , current systems are inadequate to handle the current and
Current Analysis Pipeline
future requirements needed to address bioinformatics needs .
Since 2006 , cloud computing service providers ( CCSPs ) have become publicly available and can greatly enhance NBS analysis pipelines . However , broad utilization is not a common practice . Commonly perceived barriers to adoption include data security concerns as well as a lack of expertise regarding rapidly advancing infrastructure and tools . In reality , CCSPs are one of the most secure and HIPAA-compliant technologies available , suggesting that when implemented properly they may be more secure than on-site environments . CCSPs also lower the burden on system experts as the providers maintain the infrastructure and security protection .
A Flexible Work Flow
The main advantage of CCSPs is the availability of on-demand resources to handle computational surges . This elasticity and scalability guarantee that the turn-around time advantages are independent of throughput demands . Secondly , with cloud computing infrastructure all data and resources can be centrally located and accessed remotely . This has become especially important in the context of the current COVID-19 pandemic when the workforce had to quickly adopt distributed solutions . Lastly , CCSPs provide flexibility for laboratories to move entirely to the cloud
Cloud Computing Pipeline
Figure 1 . Illustrative comparison of current and proposed cloud computing analysis pipeline . The big advantage of the proposed pipeline is that the analysis could be done in about eight hours regardless of the number of samples while the current analysis pipeline takes about 48 hours for a limited number of samples ( 2-8 ). or adopt hybrid solutions integrating with current on-site networks . With no need to entirely replace current systems , laboratories can “ ease in ” by optimizing processes that are computationally intensive and / or require remote access . In view of these benefits , the NBS community at large might benefit from integrating cloud computing into workflows .
Referencing an actual implementation strategy , Utah Public Health Laboratories will begin by automating the upload processes of raw genomic data to a cloud computing storage service every night ( Figure 1 ). Next , the data will be processed by analysis pipelines which will create a summary report . This data will be stored both on-site and redundantly on the cloud where it will be reviewed by a bioinformatician followed by a genetic counselor . The final results or reports will be generated in the cloud environment but will be accessible through the LIMS .
The big advantage of the proposed pipeline is that the analysis could be done in about eight hours regardless of the number of samples while the current analysis pipeline takes about 48 hours for a limited number of samples ( 2-8 ). But this implementation will also centralize tools , allow remote access , and significantly reduce analysis times . In the long-term , this transition not only improves analysis speed but will also reduce costs , and improve the patient experience . n
22 LAB MATTERS Winter 2021
PublicHealthLabs @ APHL APHL . org