Critical Steps in Data Management During a Crisis

Abstract

Data management is essential in a flow cytometry (FCM) shared resource laboratory (SRL) for the integrity of collected data and its long-term preservation, as described in the Cytometry publication from 2016, ISAC Flow Cytometry Shared Resource Laboratory (SRL) Best Practices (Barsky et al.: Cytometry Part A 89A(2016): 1017-1030). The SARS-CoV-2 pandemic introduced an array of challenges in the operation of SRLs. The subsequent laboratory shutdowns and access restrictions brought to the forefront well-established practices that withstood the impact of a sudden change in operations and illuminated areas that need improvement. The most significant challenges from a data management perspective were data access for remote analysis and workstation management. Notably, lessons learned from this challenge emphasize the importance of safeguarding collected data from loss in various emergencies such as fire or natural disasters where the physical hardware storing data could be directly affected. Here, we describe two data management systems that have been successful during the current emergency created by the pandemic, specifically remote access and automated data transfer. We will discuss other situations that could arise and lead to data loss or challenges in interpreting data. © 2020 International Society for Advancement of Cytometry.

Keywords: SRL; best practices; data management; flow cytometry shared resource lab; laboratory preparedness.

Figure 1

Results from the COVID‐19 data management survey with 28 respondents. Twenty‐one of the 28 respondents are in academic settings. The majority of all respondents have in‐house solutions for data management. The SRL is responsible for long‐term storage in over half of the labs that responded, with users being solely responsible in 13 of the 28 laboratories. Automated data transfer was reported in only 25% (7 out of the 28) of SRLs surveyed. Only two of the 28 respondents indicated that the pandemic impacted their standard method for data management. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 2

Workflow for adding new users: SRL staff initiate an automated script when a user fills out the request form to establish an account. When the script runs, it creates an account to grant permission to the scheduling program and send an email to the user with directions on how to use the scheduling calendar to book instrument time and training. The account established with the user's UWNetID is used to track all user processes, including managing user data, granting permission and access to software, and administrative functions like tracking training and billing. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 3

Each workstation exports collected data to a folder path labeled with the user's login ID. The data collected on any of the acquisition software platforms; Becton Dickinson FACSDiva™ software (San Jose, CA), Cytek Spectroflow® (Fremont, CA), and Amnis IDEAS (Seattle, WA) are all set up to follow a similar network path for accessing the data folders. A tree data structure is used to keep track of which files have already been uploaded to Google Drive. Whenever the script finds a new user folder, a child node of the root node is added and has its value set to the name of the folder, which in this case, is the user's UWNetID. A second child node is added under the “user node” and has its value set to the name of the instrument the first‐time user was using. Any other instruments used by that individual in the future will be added as child nodes of the “user node.” The file structure that the user creates within their folder on the instrument workstations is recreated under the corresponding “instrument node,” with nodes created for each folder and file, the file nodes being leaf nodes. Before a folder or file is added to the tree, a request is sent to GD via the GD API to create that folder or upload that file to GD. If the request is successful, the folder or file is added to the tree, so the script knows not to create/upload it again. The ID that GD assigns to the new folder/file is stored in the corresponding node and is used to tell GD where to place child folders/files in the future. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 4

Data backup and storage pathway at UT Health San Antonio. [Color figure can be viewed at wileyonlinelibrary.com]