Validation of Promega’s Casework Extraction, DNA IQ™ and PowerQuant® Systems on the Hamilton Microlab STAR Liquid Handling System
Alacoque Browne, Saileóg O’Keeffe, Kristen O’Connor, Caragh Stapleton, Sharon Murray, Dorothy Ramsbottom
Forensic Science Ireland, Garda Headquarters, Phoenix Park, Dublin 8, Ireland.
DNA profiling has become an evidentiary pillar in forensic science--used to support convictions or exonerations within the criminal justice system. The procedure by which DNA is extracted can be considered a crucial primary step in the generation of a DNA profile. This initial molecular manipulation determines both the quantity and quality of DNA yield from a forensic sample. It is therefore essential that the procedure ensures sample integrity.
Before an established method or procedure may be employed in a forensic laboratory, an internal validation must be completed to determine that the method performs as expected. This article details a direct comparison between Promega’s Casework Extraction Kit and DNA IQ™ system with the laboratory’s previous extraction system, referred to as ‘Kit X’, processed on the Hamilton STAR platform. It further presents a comparison of Promega’s PowerQuant® System with the laboratory’s previous quantification kit, referred here forward as ‘Kit Y’.
Experiments were designed to fulfill conformance with the FBI Quality Assurance Standards for DNA Casework Laboratories and the validation guidelines outlined by the Scientific Working Group on DNA Analysis Methods and ISO/IEC 17025 guidelines. Internal validation studies include sensitivity, known and mock evidence samples, reproducibility and precision, mixtures, inhibition and contamination assessment.
The data depicted here are a small representative of a much larger body of work completed by Dr. O’Keeffe, together with staff from Promega and Hamilton robotics for an internal validation at Forensic Science Ireland.
Extractions were performed in triplicate using blood from 2 donors. Volumes of blood extracted included 50 µl, 10 µl, 5 µl, 1 µl, 0.2 µl, 0.04 µl and 0.008 µl. Extractions were performed using both Kit X and Casework Extraction/DNA IQ™ and quantification data for both protocols were compared. With high volumes of blood (5-50 µl) the Kit X extraction protocol yielded marginally higher concentrations of DNA but with greater variability. The Casework Extraction/DNA IQ™ chemistry protocol, however, was more reproducible between extractions (Figure 1). Furthermore, with lower blood volumes, 1 µl and below, the Casework Extraction/DNA IQ™ protocol provided approximately 30% more DNA than the Kit X chemistry, giving higher yields with low level samples (Figure 2).
Figure 1: Sensitivity of Kit X and Casework Extraction/DNA IQ™ chemistries: DNA concentration recovered from 50 µl, 10 µl and 5 µl blood. 2 donors (Bld 1 = male, Bld 2 = female).
Figure 2: Sensitivity of Kit X and Casework Extraction/DNA IQ™ chemistries: DNA concentration recovered from 1 µl, 0.2 µl, 0.04 µl and 0.008 µl blood. 2 donors (Bld 1 = male, Bld 2 = female).
The same sample type was extracted multiple times within a run and across runs to show the repeatability and precision of the method using both Kit X and Casework Extraction/DNA IQ™ chemistries. The precision and accuracy of the extraction system was evaluated by comparing the quantification data of replicate samples to determine variance in the results. Three replicates of liquid blood (10 µl) and saliva (20 µl) from 1 donor were extracted four times: twice by the same individual on two different Casework Extraction/DNA IQ™ Hamilton runs (‘IQ1’ and ‘IQ2’). The third set of 3 replicates of blood and saliva were extracted by a second operator on the same Casework Extraction/DNA IQ™ Hamilton protocol (‘User#2’), and a fourth set of replicates was extracted using the current Kit X extraction protocol on a second Hamilton instrument (‘Kit X’). The blood extracted on the first plate (‘IQ1’) had a slightly higher DNA concentration than the 3 other extractions but overall, the performance with the blood samples shows the Casework Extraction/DNA IQ™ kit yields a higher average concentration of DNA (Figure 3).
The Saliva samples extracted on the first 2 runs (‘IQ1’ and ‘IQ2’) were of similar yield, however, the DNA extracted by the 2nd user and with Kit X had a greater DNA output than with Casework Extraction/DNA IQ™ chemistries. As the variance within these particular replicates was low, it suggests the variation is not a function of the DNA purification, rather there may have been some variation in the sample amount loaded to the swabs.
Figure 3: Reproducibility of Casework Extraction/DNA IQ™ chemistries: Average DNA concentrations for blood and saliva samples, extracted by the same user (1 and 2), a separate user (3) using Casework Extraction/DNA IQ™ and Kit X (4).
Ten known samples of single source were extracted. Five samples were of male origin, 5 were from female donors. Samples were then quantified twice with the PowerQuant® System and once with Kit Y. The operator for this DNA extraction noted a low amount of resin the in well containing sample ‘Known 5’ which correlates with the very low concentration of DNA recovered from sample ‘Known 5’ (Figure 4). This was the first run on the instrument for the validation and likely a result of user error during set up of the instrument (suggest insufficient mixing of beads and lysis buffer prior to aliquoting on the robotic deck). All nine other samples all generated high amounts of DNA (Figure 4). As expected, there was some variation in the DNA concentration achieved between donors (Figure 4). The Kit Y quantification of the autosomal fragment in these samples is typically higher than with PowerQuant® System. The Kit Y autosomal data typically suggests a higher quantification of DNA than the PowerQuant® target, this effect is highest in sample ‘Known 4’ (Figure 4). Both the PowerQuant® assay and the Kit Y assay employ targets that are multicopy in the genome. This increases the sensitivity of the assay. The copy number of these targets is variable between individuals. It is possible that the Kit Y autosomal target in the DNA of sample 4 has unusually high copy number and results in a relatively high response to the autosomal target. The two PowerQuant® assays show very little variation between the two amplifications, showing the assay is reproducible.
Figure 4: DNA quantification of 10 known samples: Small autosomal target for 2 PowerQuant® assays, and one Kit Y assay.
Six additional known samples were amplified with 1 ng of template DNA in NGM SElect™ PCR Amplification Kit based on both the PowerQuant® System and Kit Y quantification values. (The DNA profile designations for both systems are identical). For the same sample, the Kit Y quantification was routinely higher than the PowerQuant® quantification (Figure 5). The amplification of the Kit Y target was therefore lower than with the PowerQuant® set up. Consequently, the peak heights of the NGM SElect™ profiles based upon the PowerQuant® assay were consistently higher than the assays assembled based upon Kit Y data (Figures 5&6). It was determined that the input DNA amount should be reduced to 0.8ng as the profile quality was equivalent. This allowed for maintaining current Genemapper™ software thresholds and STRmix™ software settings.
Figure 5: Mean Peak heights: (in rfu) for NGMSelect™ DNA profiles amplified on the basis of quantification data from Kit Y and PowerQuant® assays. Mean peak heights on left of scale as column plot and concentration of DNA on right scale as dot plot.
Figure 6: Mean Peak heights of 17 DNA targets: in 6 known samples based on quantification data from PowerQuant® or Kit Y.
Thirty casework-type samples were extracted including samples such as cigarette butts, touch samples and other sample types that are commonly encountered in the laboratory (Figure 7). The samples were quantified with PowerQuant® System, normalized and amplified with NGM SElect™ PCR Amplification Kit. The autosomal target varied between samples, the maximum being 3.97 ng/µl, the lowest being 0.0014ng/µl. Such variation is expected in casework-type samples, particularly if low template or “Touch DNA” samples are included as part of the data set (Figure 7). The three quantification targets in PowerQuant® System can be utilised to estimate DNA quality, as well as quantity. Ratios close to 1 (Figure 7, ‘Auto/Y’ column) tell us that the larger fragment has amplified with a similar efficiency to the small fragment and indicates that there is little or no degradation of the sample. With higher ratios (e.g. CW14) the indication is that the sample is more degraded, and we might expect to see a “ski slope” in the DNA profile with the high molecular weight fragments being amplified less than the low molecular weight loci.
Figure 7: PowerQuant® System data for each of the 30 casework-type samples.
Two male/female non-differential mixture sets were tested (one set; [Male:Female] displayed here). Ratios of 1:1, 1:5, 1:25, and 1:125 were extracted in duplicate with each of the 2 mixture samples. Extracts were quantified with PowerQuant® System and the data compared. DNA IQ™ extracts were amplified using NGM SElect™ PCR Amplification Kit and the data analysed to determine performance. As expected for the Male:Female mixtures, with increasing Female proportion, the quantification of the Y marker was reduced with increasing female component. The overall yield in these samples was higher for the Kit X extraction (Figure 8), as seen previously for high levels of DNA. When the data are converted to ratios for autosomal to Y, the 2 extraction systems show the same trends and the Auto:Y ratios are very similar (Figure 9). The expected ratios for each dilution are higher than the observed ratios, which suggest the samples used for production of the DNA mixtures were not of equal DNA content; suggest the male donor sample was a richer source of DNA.
The STR data for the two mixtures extracted by DNA IQ™ were analysed to determine the number of unshared alleles in the minor component of the mixture that could be detected. All unshared alleles of the minor component were detected at 1:5 mixture ratio. At 1:25, a single drop out gave a reduction from 20 to 19 alleles detected in the minor profile, equivalent to 95% detection rate. At a mixture ratio of 1:125, a small number of alleles were detected, 3 and 2 in the duplicate amplifications, equivalent to 15 and 10% of alleles.
Figure 8: DNA yields in mixture samples: Male:Female mixtures, 1:1, 1:5, 1:25 and 1:125 extracted using DNA IQ™ and Kit X extraction protocols.
Figure 9: Ratio of Autosomal to Y markers in DNA mixtures: extracted using either DNA IQ™ or Kit X. Observed and expected values are plotted for Male: Female mixtures.
Three person mixtures were created of varying ratios of blood from 3 contributors. Duplicate swabs for the same mixture ratios were made using 15 µl blood, and each sample was extracted using the Casework Extraction/DNA IQ™ automated protocol and the laboratory’s Kit X DNA extraction protocol. Samples were quantified with PowerQuant® and the data were collated and compared.
In general, the Kit X extraction gave a higher DNA concentration, though the lower level Y target appears to be recovered in the DNA IQ™ extraction to a similar degree (Figure 10). NGM SElect™ profiles of the DNA IQ™ extracts were examined and in 11 of 16 profiles all alleles from all 3 donors were present. The mixture ratios observed generally matched those expected. The mean peak heights for the samples are very similar across the sample set, indicating that the samples amplify with a similar efficiency and reproducibility.
Figure 10: Comparison of DNA IQ™ and Kit X extractions of 3-person mixtures: Graphical representation of DNA concentrations (Auto, Deg & Y) achieved by DNA IQ™ and Kit X.
The amplification performance of the PowerQuant® System IPC can be used as a guideline to evaluate the possible presence of inhibitors in a DNA sample. The IPC was tested by quantifying a controlled DNA sample mixed with different concentrations of two known inhibitors, haematin and humic acid to determine the effect on the IPC, and subsequent amplification with NGM SElect™ PCR Amplification Kit. For haematin effect investigation, concentrations of 0µM (control), 417µM, 833µM, and 1250µM were prepared. Samples were prepared with humic acid concentrations of 0ng/µl (control), 200ng/µl, 400ng/µl, and 600ng/µl. All inhibitor samples were prepared with an approximate DNA concentration of 0.11ng/µl and quantified with PowerQuant® System. Each inhibitor sample was then amplified with NGM SElect™ PCR Amplification Kit using 10µl of sample, and the data analysed. The resulting STR profile was compared to the IPC shift from the quant to determine whether any effect on the IPC quantification cycle is informative.
Increasing haematin concentrations cause inhibition of the NGM SElect™ STR amplification (Figure 11). With no haematin, the mean peak height of the alleles was in the region of 18000 rfu. The peak height with 417 µM haematin added was reduced to around 12000 rfu. At 833 µM haematin, a single peak was observed in one duplicate with a height of 200 rfu. All other alleles dropped out of the amplification. At 1250 µM haematin no amplification was observed.
With humic acid (HA), there was little effect on the overall peak heights of the DNA profiles with concentrations of up to 200 ng/µl there was no incidence of allele drop out (Figure 12). At 400 ng/µl HA, the mean peak heights were 2100 rfu, much lower than the control, or 200 ng/µl samples, and 14 of a possible 33 alleles were found to drop out. With 600 ng/µl HA, a single allele with a peak height of 214 rfu was observed in the duplicates.
Figure 11: Mean Peak heights and allele drop out of the NGM SElect™ DNA Profiles with increasing amounts of haematin. Mean peak height (in rfu) on left side scale and number of drop out alleles on left side scale.
Figure 12: Mean Peak heights and allele drop out of the NGM SElect™ DNA Profiles with increasing amounts of Humic acid. Mean peak height (in rfu) on left side scale and number of drop out alleles on left side scale.
Multiple negative controls were included on each Casework Extraction/DNA IQ™ extraction run on the Hamilton STAR. Contamination studies were performed using both checkerboard pattern and zebra patterns (Figure 13). All negative controls were quantified using PowerQuant® System. Any sample showing a quantification value was further checked by amplification using NGM SElect™PCR Amplification Kit. Across all contamination studies, only one event was observed. One negative control sample had a low cycle threshold (Ct) for the autosomal target, and a slightly increased Ct for the Y target. Upon investigation of the real-time PCR data, the multi-component plots clearly show that no amplification is detected in Auto, Y or degradation signal in these wells, but there was a sudden shift in fluorescence detected in all channels between cycles 2 and 3 in the amplification plot. This may have been caused by a bubble in the well.
Figure 13: Checkerboard pattern (above) and zebra pattern (below) layout for contamination studies.
The DNA IQ™ System extraction protocol performed well within the bounds of this validation. Whilst the overall DNA yield of stronger DNA samples was lower than with the Kit X chemistry, the yield was more than adequate for purpose. With lower concentration samples, DNA IQ™ System provided approximately 30% more DNA than Kit X chemistry, consistently giving higher yields with low level samples. It is worth noting the DNA IQ™ beads have a maximum binding capacity designed to limit irreversible binding to the magnetic beads. As a result, lower quantifications are expected for high template samples and so limit the total amount of DNA recovered. Nevertheless, blood samples equivalent to 0.04 ng/µl DNA extracted by the DNA IQ™ System is sufficient as to allow an informative DNA profile to be generated.
Known samples extracted and quantified using DNA IQ™ and PowerQuant® Systems produced the same profile designations as the laboratory’s current system.
The DNA IQ™ and PowerQuant® Systems combination was shown to be capable of performing extraction and quantification with a number of different DNA sample types and from very variable sample quantity. The PowerQuant® System can additionally be used to predict profile quality from the quantification data.
Mixture samples extracted by the system behaved as expected. In 2 person mixtures of male:female samples, the Auto/Y at quantification correlated with the amelogenin X:Y ratio in the STR data.
With 3 person mixtures, the DNA profiles largely reflected the input blood ratios showing that the DNA IQ™ System performs as expected in the extraction of DNA from multiple donors.
No contamination was observed throughout the validation, either at quantification stage following the DNA IQ™ extraction controls, nor in the NGM SElect™ DNA profiles.
The studies performed in this validation met the criteria for an internal validation and have shown that the Casework Extraction/DNA IQ™ and PowerQuant® Systems are suitable for use in a forensic laboratory.
This validation was assessed by the Irish National Accreditation Board (INAB) and accreditation awarded.
Alacoque Browne is currently a forensic analyst with Forensic Science Ireland in Dublin, and has been with the organization since 2019. Prior to that, she was a Postdoctoral Researcher with Dublin City University. Alacoque received her Bachelor of Science degree in Physiology from the National University of Ireland, Galway, and her Master of Science in Reproductive and Developmental Biology from Imperial College London. After, she completed her PhD in breast cancer at the Royal College of Surgeons in Ireland.