About Southern Research Institute
Southern Research Institute is an Alabama-based independent, not-for-profit center for scientific research. Consisting of a diversified network of collaborative facilities, Southern Research is a recognized leader in major scientific discoveries in drug discovery and development, engineering, aerospace, materials, chemical and biological defense, and environment and energy research.
Up until 2007, Southern Research Institute’s High Throughput Screening (HTS) department used ActivityBase and its Microsoft® Excel UHTE templates to fulfill all their screening requirements. At that time the department was limited to running single dose and in-plate concentration response single endpoint assays to calculate inhibition using simple UHTE templates.
Fig 1 and Fig 2: Southern Research Institute’s fairly simple UHTE templates for single use and concentration response single point assays
As their research evolved, screening requirements changed to include single dose and in-plate concentration response kinetic assays with 10 readouts per time point, causing the Excel UHTE template to grow slightly in complexity. The new template, however, was still manageable by researchers.
Fig 3: Slightly more complex single dose kinetic assay UHTE template
When template requirements grew even further to accommodate a request for more data calculation columns, the spreadsheet expanded to populate the entire range of Excel columns from A to Z. QC requirements also added another level of complexity as new functionality to ensure quality and run-to-run consistency was added to the template.
Fig 4: The single dose kinetic assay UHTE template grows more populated as additional calculation columns are added
Median and standard deviation data had to be tracked in a ‘Historical Control’ document over the entire campaign, resulting in the addition of new charting and statistical functionality to determine whether the assay results were of a high enough quality. Median values for various sections of each plate were compared to demonstrate random distribution of compounds and also to facilitate identification of possible liquid handling inconsistencies.
In the Excel UHTE template, plate control metrics were returned from the summary sheet as non-object related data. The data was used to determine whether to verify/reject data-specific plates.
Researchers began calculating stacked plate concentration response in single endpoint assays by creating 10 stacked serially diluted copies of each compound plate to generate a concentration response curve. Running stacked plates, with multiple time points on each plate, was considerably easier using robotics.
Fig 5: Stacked plate concentration response kinetic read
However, preparing the Excel UHTE template for the stacked plate single endpoint assays involved adding 10 concentrations, each in duplicate, to populate the template through column C to M. In addition, each plate in the dilution series had to be read at 10 time points, resulting in 100 raw data points per well, without taking duplicates into account.
Excel’s ability to accommodate the number of necessary columns and worksheets when recording data in three dimensions was restricted, resulting in a highly complex template which proved difficult to maintain.
Using ActivityBase XE
In late 2007, researchers at Southern Research Institute began using ActivityBase XE for their kinetic primary and concentration response screening assays, beginning with a slightly modified version of the kinetic protocol and a single dose screen.
Fig 6: Modified single dose kinetic protocol in ActivityBase XE
Using XE Designer, the template building component of ActivityBase XE, the following slope calculation was added to the template, returning the gradient of a simple linear relationship between two given data arrays:
- x Numeric List
- y Numeric List
- xy x*y
- size Size(xy)
- slope (Sum(xy)*size-Sum(y)*Sum(x))/(Sum(x*x)*size-Sum(x)*Sum(x))
Returns: Numeric List
After adding the Slope calculation to the XE template, the scientists were able to easily add Rate and Percent Inhibition calculations as well. The concentration response protocol was also slightly modified from the UHTE version by adding intra-plate QC calculations. A virtual plate was added to enable the template to accommodate cross-plate screening.
Fig 7: Modified concentration response kinetic protocol in ActivityBase XE
Viewing the Results
After the modified templates were created in XE Designer, the assays were performed in XE Runner, where the Dose Response Table view was used to look at results, including dose response curves, curve status and inhibition values for each concentration at each measured time point.
Fig 8: Viewing results in ActivityBase XE Runner’s dose response table
Researchers found the learning curve a bit steep at first when moving from Excel to XE templates in terms of calculations, which were initially considered to be fairly abstract in ActivityBase XE, requiring a more coded approach. However, instead of Excel’s highly complex template involving 10 or more worksheets, XE easily accommodated all the necessary dimensions for both the single dose and concentration response assays, so offering immediate benefits in terms of manageability.
It was also found to be straightforward to implement QC controls when the scientists discovered that they were able to create the same plate control calculations in the ActivityBase XE environment as they used in Excel, and could easily continue their strategy for identifying questionable plates. The new calculation format in XE allowed the researchers to continue viewing their Historical Controls in the same way as in Excel.
In addition, the screeners discovered new views and tools in ActivityBase XE that could be used for additional QC, such as the Tiled Plate View which shows up to 35 by 35 plates at any one time. Scientists could scroll through potentially thousands of plates, displaying color-coded wells based on any numeric value, to quickly spot areas of activity or potentially erroneous results and identify trends within and across the plates.
Fig 9: ActivityBase XE’s tiled plate view
Wells could be selected and excluded from this view, and once excluded all calculations that were dependent on that well value, such as average, standard deviation or inhibition values, were automatically recalculated. Wells could also be excluded from a series of plates, so for example if there was a pipetting problem with one or a set of wells on every plate, these could be excluded by selecting the wells and applying the exclusion to every plate.
The Aggregate Plate View was also used to detect a well that may be a ‘hot-spot’ of activity. Using this view, any aggregate statistic can be calculated for any piece of well-based data just by selecting the wells and choosing the statistic, for example, Average, from a list. Wells can also be excluded in this view, so any wells that contribute to an aggregate statistic can be invalidated and color-coded to indicate exclusion.
Fig 10: Aggregate plate view
The Cross Plate Chart, a useful view for looking at plate-based calculations across an entire test, provided Southern Research Institute’s screeners with a run-level view of the data tracked over the full campaign using their Historical Controls.
Fig 11: Cross plate chart
Scientists also liked the Cumulative Distribution view which displays control/sample ranges, with an export function that provides a quick way to retrieve a list of ‘hits’. Cumulative Distribution also offers a means of bulk rejection of statistical outliers.
Fig 12: Cumulative distribution
The Frequency Distribution view displaying inhibition values was also valuable to Southern Research Institute’s QC processes, while Test Well Scatter offered a different view of control/sample anges of interest, and also allowed researchers to export/reject data points.
Fig 13: Frequency distribution
When running straightforward single endpoint assays using ActivityBase and UHTE templates, screening requirements at Southern Research Institute’s HTS department were fulfilled. However, as more complex kinetics assays that generated data for multiple time points were introduced, Excel was not able to accommodate the increased data volume. In addition, applying and managing QC processes resulted in a hugely complicated template that was almost impossible to maintain.
By moving to ActivityBase XE, screeners were able to easily create new templates with built-in functionality that accommodated their multi-dimensional kinetic assay data. With the ability to support an unlimited number of time series data points, ActivityBase XE provided Southern Research Institute with a user-friendly environment for template building, results viewing and QC control.
In addition, ActivityBase XE delivered a number of visualization tools that made the whole QC process both easier and more efficient, allowing Southern Research Institute to apply greater consistency to their screening data and improve the quality of results generated.