Figure 1.
The three LRE Analyzer databases.
The experiment database holds a group of related runs, a concept taken from the RDML guidelines for exchange of qPCR data [8]. Similarly, the calibration database holds calibration profiles used for optical calibration, whereas the amplicon database contains amplicon information. Data is organized into tree-like structures, which provides a convenient method for viewing and editing data. For the experiment database, each run is presented as a branch under which the profiles generated during each run are listed. As described in the text, the replicate sample profiles are used to generate an average sample profile. Similarly, the calibration database holds calibration profiles organized under a reaction setup, from which an average optical calibration profile (OCF) is generated. The primary function of the amplicon database is to provide amplicon sizes during run import, which in combination with an average OCF is used to determine the number of target molecules within each sample.
Figure 2.
The user interface is organized into three panels.
The explorer panel contains windows for viewing data within each of the three LRE database types (Figure 1). The central editor panel contains windows for viewing/editing of profiles and amplicons. The sorting panel allows viewing of profiles generated over multiple runs, sorted by either amplicon or sample. Note that the two sorting windows can be docked onto the left hand border of the main window, allowing the main window to be resized. This provides a convenient method for accessing the sorting windows while reducing the overall size of the main window. Note also that changes to the main window are saved across sessions so that the main window is restored to its previous state when the program is restarted.
Figure 3.
The explorer panel contains three windows for viewing of data within the three LRE databases.
In this screenshot, the demonstration experiment database has been opened, which contains data generated by three runs conducted over 3 days, with the first run expanded to show nine average sample profiles. The sample profile labels are generated using the template: <“amplicon name” @ ”sample name” (Emax) # target molecules>, and illustrate the primary output of the LRE Analyzer, which is the automated determination of the number of target molecules within a sample. In this example, transcript quantities have been determined for three genes (12240, 46630 and GAPDH) within three replicate reverse transcriptase reactions (COL1-3).
Figure 4.
The editing panel contains windows for editing profiles and amplicons.
In this example, a replicate sample profile has been selected in the experiment database explorer window, which triggers a display of information associated with this profile. Although a description of each of the subpanels will not be presented here, the LRE Analyzer help set provides a detailed description of how each one functions, along with how the number of target molecules is determined, which in this example is 1,909 molecules.
Figure 5.
The sorting panel allows profiles to be organized by either amplicon or sample.
These windows become active whenever the experiment or calibration explorer window is active. Selection of either an amplicon or sample (an amplicon in this example) will generate a list of all profiles within the database (an experiment database in this example) generated with that amplicon or sample. This in turn allows profiles generated across multiple runs to be viewed/edited or exported as a group. Note that selecting the “Run View” button within the experiment explorer window will restore the run-based tree view. Note also that the sorting windows can be iconized using the button located in the upper right side of each window. Placing the mouse over either of the iconized windows will trigger the respective window to fly out, allowing an item to be selected. Selecting an item in the explorer window will then trigger retraction of the sorting window, providing a convenient method for accessing the sorting windows while reducing the overall size of the main window.
Figure 6.
An Excel template provided by the LRE Analyzer for manual import of sample profiles.
A similar import template is provided for calibration profiles. Raw fluorescence readings for each replicate profile are pasted into the template, along with amplicon and sample name, amplicon size and the strandedness of the target, for each replicate profile. The Excel workbook is then selected within the LRE Analyzer to initiate profile import. Note that details about data import and export are provided within the LRE Analyzer help set.
Figure 7.
An example of data export sorted by amplicon,
in which profiles generated by each amplicon are placed into a separate worksheet. Data sorted by run or sample are similarly exported. No: the number of target molecules; C1/2: the fractional cycle at which reaction fluorescence reaches half of maximum (similar to, but more reliable than, Cq).
Figure 8.
The LRE window encompasses a contiguous group of cycles that are used for LRE analysis of a profile, depicted as red and black circles within the FC and LRE plots, respectively. Linear regression analysis is then used to determine values for the two parameters upon which LRE analysis is based (Emax from the Y-intercept and ΔE from the slope). Although LRE window selection is fully automated, the program does allow manual adjustment of the LRE window using the buttons within the LRE plot.
Figure 9.
The LRE window selection parameters panel.
A. Default settings in which the start cycle is set to the first cycle below C1/2 and the F0 threshold set to 6.0%, which is used to determine the top of the LRE window (see the text for details). This produces an average replicate F0 CV (Av Repl-Fo CV) of 29.0%, which is a general indicator of intra-run variance generated by LRE analysis (see the text for details). B. The minimum FC has been manually set to 225,000 fluorescence units, such that the start cycle is set to the cycle following the first cycle that generates a fluorescence reading above this minimum FC. This reduces the average replicate F0 CV to 25.2%. Note that the LRE Analyzer help set provides additional details about LRE window selection.
Figure 10.
A common form of kinetic distortion, referred to as “plateau drift”, is produced by a continued increase in amplicon DNA quantity beyond that predicted by the LRE model, which is represented as circles within the upper panel (referred to as the FC plot). This is particularly evident in the LRE plot (lower panel) as a progressive drifting of points above the LRE line. Importantly, inclusion of these aberrant cycles in the LRE window (represented by the black circles in the LRE plot and the red circles in the FC plot) will generate an underestimation of Emax (Y-intercept) that leads to an overestimation of target quantity.
Figure 11.
The tabular summary provides values for the three parameters that define a cycle.
C: cycle number, FC: the fluorescence reading, EC: the cycle's amplification efficiency, %Av. Fo: the percent difference between the cycle F0 and the average F0 generated by the cycles within the LRE window (designated by the red font). In this example, the F0 threshold (Figure 9) was set to 6% so that cycle 29, which generated a 6.84% difference, triggered termination of LRE window expansion at cycle 28.
Figure 12.
An example of profile collapse characterized by progressive drifting of points below the LRE line.
In this example the collapse is produced by the primer pair, although reduced enzymatic activity, such as that produced at high concentrations of SYBR Green I, have also been found to generate profile collapse for primer pairs that normally conform well to the LRE model. Similar to plateau drift (Figure 10), it is important to exclude such aberrant cycles from the LRE window, which for profile collapse will lead to an overestimation of Emax (Y-intercept), which will generate an underestimation of target quantity.
Figure 13.
An example of extreme profile arcing produced by the enzyme formulation.
Similar to plateau drifting (Figure 10), inclusion of these aberrant cycles into the LRE window will generate underestimations of Emax (Y-intercept).
Figure 14.
Excel summary of the cDNA quantifications contained within the demonstration experiment database.
Expressed as the number of transcripts per 2.5 ng of total RNA, these datasets provide insights into both the utility of absolute quantification and the precision that can be generated by the LRE Analyzer. This reveals a run-to-run CV of ±5–16% for all three targets, and a RT-to-RT CV of ±1–20%. This summary also presents LDA quantifications that provide an independent determination of target quantity, with the difference expressed as a percentage (%Diff). Note that the GAPDH LDA failed due to the frequent production of non-specific products when the target quantity was diluted below one molecule per aliquot.