CytoSure Services
Sample Requirements and QC Metrics for CytoSure™ Services when using CytoSure ISCA (8x60k) arrays with CytoSure Labelling Kits
Introduction
The most important step in achieving good quality data is making sure that your samples arrive at OGT in perfect condition. The first section of this article describes how to ship your samples and how the samples should be ordered within the plate. To ensure that you obtain the best quality data, please endeavour to achieve the DNA quality values we recommend. On arrival at OGT, the first QC step is performed to confirm the quality of your DNA. If there are any problems or questions arising from this one of our technical experts will contact you to discuss any issues in more detail. If needed at this stage, we can offer technical advice on how to improve the quality of your samples. Please contact us if you would like advice regarding DNA extraction or performing your own QC checks.
As the array processing procedure continues many more QC checks are made and these checks are described in detail below. At each stage if your samples do not perform as expected we contact you to discuss whether to continue with the experiment. These contact points are described in this article for each of the QC metrics. It is important to us that you are kept informed regarding the status of your experiment and that you are confident in the quality of data you receive.
DNA QC and Shipping
How to ship your samples:
-
Samples should be suspended in TE 50 ng/µl (see ’Required DNA QC Metrics’ for more detailed information)
-
Use a 96-Well Eppendorf Twin Tec Plate (p/n 0030 132.513)
-
Seal with ABgene 12-Domed-Cap Strips (cat. no. AB-0853) or with Thermo Scientific Easy Peel Heat Seals (p/n AB-0745)
-
Do not use adhesive plate seals
-
Pack with sufficient dry ice to last the delivery time
-
For example, an insulated 19 litre box would require 11kg of dry ice pellets for shipping in Europe
-
Protect the plate by wrapping in cardboard
-
The samples will be unpacked upon delivery.
If there is no dry ice left in the box, you will be informed and a decision will be made whether to continue as the quality of your samples could have been compromised.
How to deposit your samples in the plate:
-
Pipette samples in wells E1-H11
-
Pipette sex matched references in wells A1-D11
-
Plate layout is shown below:
Figure 2: Plate Layout
After labelling the samples and references will be combined as follows:
-
Test samples from row H are added to reference samples in row D
-
Test samples in row G are added to reference samples in row C
-
This is continued for each row of test samples as shown in Figure 3 below
-
Sample and reference are always combined in the same order, sample H1 is added to D1, H2 to D2 and this is continued along the row
-
Well D12 always contains the OGT internal control (Human Genomic DNA Promega, cat. nos. G1521 [female] or G1471 [male]). This sample is referred to in this article as ‘the internal control sample’
Figure 3: Combining reference and test samples
Required DNA QC metrics:
-
Samples are measured on arrival
-
You will be contacted if the QC metrics fail
-
A decision will be made whether to proceed
-
If the samples fail QC, OGT cannot be held responsible for subsequent poor data unless caused by array manufacturing or processing failure
DNA QC metrics: (8 x 60k)
Volume: >20 µl
Concentration: Exactly 50 ng/µl
A260/230: >1.5
A260/280: >1.8
Test and reference samples must pass all of these criteria.
Labelling QC
For all labelling reactions a control sample is processed alongside your samples.
After labelling:
-
Sample concentration is measured
-
Amount of dye incorporated is determined
Labelling QC metrics:
Yield: >10 µg
Specific Activity: >15 pmol/µg
Values less than this, in either channel, for test, reference and control samples are considered a failure.
The diagram below illustrates the decision making process used to decide the next step after determining the labelling QC values:
Figure 4: Decision tree for Labelling QC
*Note: If less than 10% of the samples fail labelling QC, the array process continues. The format of the array (8 pack) dictates that all arrays have to be used, an empty array cannot be used at a later date. Therefore, it is preferable to continue to process a small number of failed samples rather than leaving an array unhybridised.
†Note: If the control sample fails, it is assumed that the fault of failure is due to a technical error and the labelling reactions are repeated at OGT’s expense.
Array QC Metrics
Derivative Log Ratio Spread (DLRS)
This metric measures the variation of the data, which is primarily related to sample quality. It is calculated by looking at neighbouring oligonucleotides and determining the variation in log2 ratio along a chromosome.
The following values are considered failures for DLRS:
Internal control sample fail: >0.20
Test sample fail: >0.30
The diagram below illustrates the decision making process used to decide the next step after determining the DLRS values:
Figure 5: Decision tree for DLRS Values
*Note: if the DLRS value is >0.3 this is likely to be due to a sample problem and you will be contacted by one of our experts to discuss ways of improving the data quality.
Background Noise in Red and Green
This metric is calculated as the standard deviation of negative control probes on the array. The values are recorded and can be classified into the categories below. A poor background does not necessarily indicate that the array has failed. This is a secondary metric as it is incorporated into the Signal-to-Noise metric (see below).
Green
Excellent: <15
Good: 15–20
Poor: >20
Red
Excellent: <25
Good: 25–35
Poor: >35
Signal Intensity Red and Green
This metric measures the signal intensity on the array. The values are recorded and can be classified into the categories below. Poor signal intensity does not necessarily indicate that the array has failed. This is a secondary metric as it is incorporated into the Signal-to-Noise metric.
Excellent: >1000
Good: 800–1000
Poor: <800
Signal-to-Noise Red and Green
This metric combines the background intensity and the signal intensity metrics into a single value. It is a good indicator of array quality. The values can be classified into the following categories:
Excellent: >100
Good: 30–100
Poor: <30
A value of less than 30 for test samples or control samples is considered a failure.
The diagram below illustrates the decision making process used to decide the next step after the signal-to-noise values have been determined:
Figure 6: Decision tree for Signal-to-Noise metrics
*Note: If an array fails the signal to noise metric but all other metrics (excluding reproducibility [see below]) pass, you will be given the option to repeat the array. If the repeat array passes the signal-to-noise metric, OGT will pay for the repeat test. If the sample fails the repeat test, you will pay for both arrays. The assumption will be made that it is a sample problem if it fails twice.
Reproducibility Red and Green
This metric is used to measure technical variability across the array. Likely causes of failure are poor mixing during hybridisation caused by:
-
Oven failure
-
The bubble formed on setting up the hybridisation chamber failing to move freely
-
Hybridisation solution leaking from the gasket slide.
The values for this metric can be divided into the following categories:
Excellent: <0.10
Good: 0.10–0.20
Poor: >0.20
A test sample or control sample has failed if it has a values of >0.30.
The diagram below illustrates the decision making process used to decide the next step after the reproducibility metrics have been determined:
Figure 7: Decision tree for Reproducibility metrics
Feature Non-Uniformity Outliers
A feature is flagged as a non-uniformity outlier when the signal intensities of the individual pixels within that feature do not follow a normal distribution. A skewed or bimodal distribution of pixel intensity values can indicate a processing or manufacturing problem.
A test or control sample has failed if there are more than 1% feature non-uniformity outliers in either channel.
The diagram below illustrates the decision making process used to decide the next step after the feature non-uniformity outlier values have been determined:
Figure 8: Decision tree for Feature non-uniformity outlier metrics
*Note: It is very rare that a high number of feature non-uniformity outliers occurs. If this is not related to a manufacturing or technical fault then you are contacted to discuss whether this is related to sample quality.
Waviness Metric
For a typical sample, in general the copy number is unchanged, as shown using a log2 ratio between the test and reference samples of zero. For some samples, this ratio is not zero but fluctuates around zero, along the chromosome, generating a ‘wavy’ profile. A sample with a high waviness metric may be difficult to analyse. The values for this metric are reported and can be classified into the following categories:
Excellent: <0.025
Good: 0.025–0.027
5Poor: >0.0275
A test sample or control sample has failed if it has a values of >0.0275.
The diagram below illustrates the decision making process used to decide the next step after the waviness metric has been determined:
Figure 9: Decision tree for waviness metrics
Call Rate
OGT will report the call rate; this is the number of aberrations identified by the analysis software (CytoSure Interpret Software) in an individual sample after running the aberration detection algorithm. This is dependent on the sample and how the analysis is carried out, and therefore thresholds cannot be set for this metric.
Visual Examination of the Tiff Images
All the tiff images of the arrays are examined after scanning. This is subjective but is performed by experienced personnel. It is intended that any processing or manufacturing problems that can be detected by visual inspection are identified at this stage.
Discover more about CytoSure Services or contact us to discuss your requirements
CytoSure™ products are for research use only
Find out more
For more information about any OGT product or service, please contact us.