Assay quality combines metrics that allow you to track quality across one or multiple assays and across different time periods. All these metrics can be filtered by time period & assay(s) in the top bar:
The average value of each parameter is visualized in bold for the data. Below this average, the trend is calculated and marked green (improvement) or red (regression). The trend is calculated for the previous period of the same length as selected in the Metrics filter. Click on 'see detail' to zoom in on the data for each parameter.
Assay performance
These parameters are tracked in the Assay performance window:
Call rate
Call rate is the % of regular datapoints (all datapoints except for controls) with a call (XX, XY, YY) relative to all regular datapoints (XX, XY, YY, uncalled, inconclusive, shortfall, bad).
NTC Validity
NTC validity is the % of passed NTCs out of all NTCs.
PC Validity
PC validity is the % of passed positive controls (or passed reference samples) out of all PCs or references.
Shortfall rate
Shortfall rate is the % of shortfalls out of all samples (regular and other types).
Assay automation
These parameters are tracked in the Assay automation window:
AI Call rate
AI Call rate is the % of regular datapoints (all datapoints except for controls) that were called by the algorithm (XX, XY, YY) relative to all regular datapoints (XX, XY, YY, uncalled, inconclusive, shortfall, bad).
Manual overrides
Manual overrides is the % of datapoints for which the algorithm call was changed by a user. This metric is split into:
Manual completions: a user edited from uncalled to called (XX, XY, YY)
Manual corrections: a user edited a call to a different call (eg. XX to XY) OR a user edited from a call to uncalled, inconclusive or bad.
Best & worst performing assays
The assay leaderboard highlights the best and the worst performing assays, offering insights into assay quality and performance. These tables can be sorted by several key indicators that are correlated to the quality of the assay:
Call rate: the % of regular datapoints (all datapoints except for controls) with a call (XX, XY, YY) compared to all regular datapoints (XX, XY, YY, uncalled, inconclusive, shortfall, bad).
AI call rate: the % of regular datapoints (all datapoints except for controls) that were called by the algorithm (XX, XY, YY) relative to all regular datapoints (XX, XY, YY, uncalled, inconclusive, shortfall, bad).
Corrected data points: the number of datapoints for which the call was changed from one cluster to another cluster (eg. XX to YY) by the user.
Completed data points: the number of datapoints for which the call was changed by the user from uncalled to a cluster.
PC: Positive control validity, the percentage of positive controls where the observed call matches the expected call for the given control.
NTC: NTC validity, the percentage of negative controls (NTCs) that were correctly left uncalled due to the absence of fluorescence signal.
In addition, the tables display usage statistics that reflect the frequency and recency of assay utilization:
Frequency: a measure for the % of analyses this assay was used in.
Datapoints: The total number of datapoints the assay.
First / Last used: timestamps indicating when the assay was first and most recently used.
Read selection
The read selection shows how many times a read was chosen out of all available reads. For each read number that's available in the selection, a vertical bar visualizes the total times this read was available and the % when this read number was actually used.
As Genotyper selects a read per assay, the read selection metric represents assay-PCR plates. This is the combination of a physical PCR plate on which datapoints are tested for an assay; 1 physical PCR plate can be split into multiple assay-PCR plates.
The read selection metric can help your lab determine to optimal number of reads to be executed for a given set of assays.