Hits

Hit selection tool

Overview

The Hit selection tool allows extraction of wells that have their parameter values above a given threshold. Parameter values are normalized to that of a Negative and Positive control present in the layout.
This tool can be explored using the Screening and Screening2 examples. See the [Examples] page.
The tool loads the layout and result files from PlateEditor and allow the selection of wells that will match with the selected threshold. The method to use for normalization of the parameter values can be adjusted using the options ribbon at the top of the page.

Normalization method (1)
This option will guide whether the normalization process is done at the Global or Plate level, or use fixed, user-entered values instead of the control values:
- Global: in this mode, the entire result file is scanned to collect and average the control values before running the hit selection process.
- Plate: in this mode, the normalization is done at the individual result plate level. Plates are processed one-by-one, independently of each-other. This mode will take longer time to complete, but may help reducing bias due to plate effects.
- Custom: this mode allows the user to specify the values to use for the normalization. Note that all parameters will share the same values, so this mode is not recommended if the parameters are not the result of the same measurement method.
In all cases, the formula used the calculate the normalization value n is as follows:

where Avg100% and Avg0% are the average values for the 100% and 0% control, respectively, and v is the parameter value.

Control High and Control Low (2)
These options allow the selection of a control area to play the role of control High/Low.
- The control said "High" is used as the 100% reference for the normalization. If you are looking at percentages of inhibition, then the inhibitor is usually the correct choice here. If you are looking at percentage of growth, then the untreated control would be better suited.
- The control said "Low" is used as the 0% reference for the normalization. Using the same examples as above, the untreated control could be the 0% inhibition reference, while the growth inhibitor would play the 0% reference for a culture experiment.
Not surprisingly, the Hit selection tool will not run if the same control is chosen as both the 100% and 0% reference. Additionally, if the 100% and 0% average values are identical, the hit selection process will be unable to compute the normalization and will display a failure note for the corresponding parameter/plate.

Threshold (3a)
This field is used to indicate the percentage cut-off for the selection of wells. Note that the inequality used is inclusive: all values superior or equal to the indicated value will be selected.
Negative values and values above 100% are accepted here, allowing the selection of outliers.

Hit limit (3b)
This control is used to prevent the hit list from overflowing the memory and crash the browser.
Using a threshold too low or selecting the wrong controls may rapidly lead to situations where most of the values in the file are eligible to be selected as hits.
This control is here to ensure the hit selection process will safely stop before freezing the page crash the browser.

Control statistics and hits reporting

Control statistics
When running the hit selection tool using the Global aggregation mode, the average, standard deviation (SD) and number of values found for the selected control will be presented in a table, as in (1).
For the plate aggregation mode, the summary will be presented as a distinct table for each control, including the average, SD and number of values found at the plate level, as in (2).

Hit reporting
Hits identified will be reported with their original plate and well location, along with the parameter value from the file (Raw value) and the computed normalized value, as in (3).
For the Global normalization mode, hits are added to the table as the file is parsed. For the Plate mode, plates are parsed one-by-one, as in (4). The process can be stopped if desired (4a), without losing the hit information for the plates already computed.
The well content will also be indicated. If the well contains one or more [Ranges] with valid [Definitions] and [Pairing] to the corresponding result plate, the paired name(s) will be resolved after the hit selection is complete. This process can also be stopped if needed (4b).

Additional notes

Parsing times
The hit selection tool needs to traverse the result file several times in order to gather the desired hits.
In the global normalization mode, the result file needs to be fully parsed twice, a first pass gather the control values, a second pass calculate the normalization and gather the hits as they come.
In the plate normalization mode, the file needs to be fully parsed once for each plate, so it will take longer to complete as the number of plate increases.
In both cases, if there are hits with [Ranges] that need to be resolved, additional parsing of the definition file will be required. To optimize this last step, the hit selection tool will gather the definition at the plate level and resolve the names of all hits belonging to this plate at once.