Abstract
Background and objective
Wilson disease (WD) is an autosomal recessive copper metabolic disorder caused by mutations in ATP7B. Sanger sequencing is currently used for ATP7B variant identification. However, the ATP7B gene contains 21 exons, which makes sequencing of the entire gene both complex and time‐consuming. Therefore, a simpler assay is urgently needed.
Methods
We performed a laboratory and clinical evaluation of an oligonucleotide microarray for the detection of 24 ATP7B recurrent mutations (except p.P992L) in Chinese patients with WD.
Results
The accuracy of the microarray was evaluated by screening for ATP7B mutations in 126 patients including 106 suspected WD samples and 20 patients with other liver diseases as negative control. Results were confirmed by Sanger sequencing. We established a reliable microarray system for the rapid detection of the 24 ATP7B mutations, with a sensitivity of 30 ng/test genomic DNA and specificity of 100% for all loci; the coefficient of variation in repeatability tests was <10%. Clinical evaluation showed an overall concordance between the microarray detection and sequencing of 100%, and 81.13% (86/106) of suspected WD cases showed ATP7B mutations by microarray detection. Microarray and Sanger sequencing identified p.R778L (50.94%), p.A874V (17.92%), p.P992L (11.32%), p.V1106I (11.32%), and p.I1148T (6.60%) as the most common mutations in WD patients.
Conclusions
Our microarray system is customizable and easily used for high‐throughput detection of certain recurrent ATP7B mutations, providing a simpler method suitable for WD genetic diagnosis in China.
Keywords: ATP7B, DNA microarray, mutation detection, Wilson disease
We established a reliable microarray system for the rapid detection of the 24 ATP7B mutations, with a specificity of 100%, and the coefficient of variation in repeatability tests was <10%. We identified p.R778L (50.94%), p.A874V (17.92%), p.P992L (11.32%), p.V1106I (11.32%), and p.I1148T (6.60%) as the most common mutations in Chinese WD patients.
1. INTRODUCTION
Wilson disease (WD), also called hepatolenticular degeneration syndrome, is an autosomal recessive disorder of copper metabolism, with a worldwide prevalence of 0.5–3/100,000. 1 Disease‐causing variants in ATP7B (which encodes the copper‐transporting P‐type ATPase) lead to protein dysfunction and copper ion accumulation in organs, especially in the liver. 2 The European Association for the Study of the Liver clinical practice guidelines for WD, released in 2012, show that if the mutations were detected on both chromosomes, the Leipzig score would ≥4, which can meet the WD diagnostic criteria. 3 However, in China, clinical guidelines for the diagnosis of WD with hepatic phenotypes have not been defined. The Leipzig score was applied currently for the diagnosis of WD in China. Therefore, the detection of biallelic ATP7B pathogenic variants is required for the diagnosis of WD, especially in patients with atypical clinical symptoms. To date, more than 1000 ATP7B gene variants have been reported (http://www.hgmd.cf.ac.uk/ac/index.php). Furthermore, the spectrum of ATP7B mutations varies between China and Western countries. 4 Our previous research identified p.R778L (c.2333G>T) and p.A874V (c.2621C>T) as the two most prevalent ATP7B variants in Chinese patients with WD. 5 Other studies in Chinese patients with WD reported recurrent ATP7B mutations with varying prevalence rates, such as p.I1148T (3.88%–13.24%), p.V1106I (0.21%–9.09%), p.T935M (1.30%–6.09%), and p.N1270S (1.94%–5.88%). 5 , 6 , 7 , 8 , 9 , 10 However, in Western countries, p.H1069Q (c.3207C>G) is the most common ATP7B variant. 6 Most variants in patients with WD are hotspot variants, and therefore, the detection of common ATP7B variants will be helpful for rapid clinical diagnosis.
Various approaches, including Sanger sequencing and next‐generation sequencing technology, have been used to detect ATP7B gene mutation. However, the ATP7B gene contains 21 exons, and these approaches have the disadvantages of being time‐consuming and/or cost prohibitive, making them unsuitable for rapid clinical genetic screens. Therefore, novel and faster diagnostic methods are urgently needed.
In this study, we designed an efficient and accurate microarray assay to detect hotspot mutations in ATP7B. The WD microarray included the 24 most common ATP7B mutant loci, covering up to 70% of Chinese patients with WD. Clinical evaluation of the microarray in Chinese patients with WD demonstrated that this microarray assay could increase the speed of screening potential WD patients. Our results indicate the potential utility of the WD microarray in clinical diagnosis laboratories.
2. MATERIALS AND METHODS
2.1. Patients and samples
We recruited 139 patients from the China Registry of Genetic/Metabolic Liver Diseases Group, including 33 WD patients (Table S1) with the 24 most common ATP7B mutations for the determination of the cut‐off value and 106 patients (Table S2) with suspected WD for clinical evaluation. Suspected WD was defined by the presence of key features of WD including liver disease and cirrhosis, neuropsychiatric disturbances, Kayser‐Fleischer rings in Descemet's membrane of the cornea, and acute hemolytic attacks that are usually associated with acute liver failure. 3 In addition, 20 patients (Table S3) with other liver diseases but without evidence of WD were included as a negative control group. Peripheral blood was collected from patients and genomic DNA was extracted for ATP7B genotyping. All patients provided written informed consent. This study was approved by the Ethics Committee of Beijing Friendship Hospital, Capital Medical University (Number 2019‐P2‐217‐02).
2.2. Primers and probes
The DNA sequence of ATP7B was downloaded from GenBank (http://www.ncbi.nlm.nih.gov/genomes/). Primers and probes were designed by software DNAMAN (https://www.lynnon.com/dnaman.html) and Primer Premier (http://www.premierbiosoft.com/primerdesign/), according to the location of the detected sites. Forward and reverse primers were chemically synthesized by Tsingke Biological Technology and the 5′ end of the reverse primer was labeled with the Cy3 fluorophore. The PCR product lengths varied between 84 and 286 bp (Table S4).
Probes were the reverse complement of target sequences. For each mutation site, four probes were designed (A, T, C, and G at the mutation site). Twelve thymine residues were added to the 3′ end of the probe to improve its flexibility, and the amino group through modification of the probes condensed with the aldehyde group and fixed the probe to the surface of the chip.
2.3. Plasmid construction
For optimization and quality control of the microarray system, three ATP7B DNA fragments containing the 24 mutated sites were chemically synthesized (Table S5) and ligated into the PGM‐T vector (Tsingke). The primer‐targeted sequences for multiplex asymmetric PCR were included in each fragment.
2.4. Microarray preparation
Probes were dissolved to 50 μmol/L in RNase‐free water and mixed with an equal volume of spotting buffer (6 × sodium citrate buffer (SSC), 0.1% sodium dodecyl sulfate (SDS)). Probes were then spotted on the surface of the aldehyde group–based microarray slide (CapitalBio) using a PersonalArrayer™ 16 spotting instrument (CapitalBio). The chip contains 10 reaction zones, and 96 probes were designed as an 8 × 12 matrix in each reaction zone to detect 24 loci. The four probes corresponding to each variant were arranged horizontally in the matrix, in the order of A, T, C and G, in duplicate. The positive control (mixture of all probes), negative control 1 (spotting buffer), negative control 2 (RNase‐free water) and indicator probes (probes of T12 oligonucleotides with Cy3 fluorophore (5′ end) and amino group modification (3′ end)) were spotted at the bottom of the matrix. Genotyping and the arrangement of the microarray are shown in Figure 1. The slides were dried for 12 h and stored at 4°C until hybridization.
FIGURE 1.
Schematic representation of the 24 ATP7B loci and the arrangement of probes in the WD microarray.
2.5. Multiplex asymmetric polymerase chain reaction (PCR)
Genomic DNA was amplified using three independent multiplex asymmetric PCR reactions (Platinum Multiplex PCR Master Mix, Thermo Fisher). The PCR components are listed in Table S6. Multiplex asymmetric PCR was performed using the following parameters: pre‐denaturing at 95°C for 2 min, followed by 40 cycles of denaturing at 98°C for 20 s and annealing extension at 65°C for 90 s, and a final extension at 72°C for 5 min. The products were denatured at 98°C for 5 min and then incubated at 4°C for 10 min in the dark.
2.6. Microarray detection
Polymerase chain reaction products were mixed with hybridization reaction buffer (5× SSC, 0.5% SDS) at a ratio of 1:4. The hybridization mixture (10 μl) was added to the microarray reaction zone and uniformly distributed. The microarray slide was protected against exposure to light, placed in a hybridization container, and hybridized in a 60°C water bath for 1 h. The slide was then transferred into pre‐warmed (42°C) washing buffer A (1x SSC, 0.2% SDS) and then washing buffer B (0.2% SSC). The slide was centrifuged at 1000 rpm for 1 min and dried at room temperature. Hybridization images were obtained using the Molecular Devices GenePix 4000B chip scanner (CapitalBio) at 532 nm.
2.7. Statistical analysis
The signal value of each probe was analyzed using ArrayVision 7.0 software. Ratios of the signal value of wild‐type probes divided by the signal value of mutant probes as well as the coefficient of variation were calculated to determine the cut‐off value for the judgement of each mutation. All statistical analyses were performed using SPSS 16.0 software (SPSS Inc.).
3. RESULTS
3.1. Design and preparation of the DNA microarray
We designed the DNA microarray containing 96 probes to detect the 24 common recurrent ATP7B mutations in Chinese patients with WD. The 24 mutations cover more than 70% of WD patients in Chinese clinical practice (Table 1). The mutations and prevalence rates include p.R778L (18.93%–45.59%), p.A874V (0.49%–18.83%), p.I1148T (3.88%–13.24%), p.V1106I (0.21%–9.09%), p.T935M (1.30%–6.09%), and p.N1270S (1.94%–5.88%). 5 , 6 , 7 , 8 , 9 , 10
TABLE 1.
List of ATP7B mutations selected for microarray detection.
| No. | Mutation a.a | Codon change | Mutation nt | Location |
|---|---|---|---|---|
| 1 | p.R778L/Q | CGG>CTG/CAG | c.2333G>G | Exon 8 |
| 2 | p.A874V | GCG>GTG | c.2621C>T | Exon 11 |
| 3 | p.I1148T | ATT>ACT | c.3443 T>C | Exon 16 |
| 4 | p.V1106I | GTC>ATC | c.3316G>A | Exon 15 |
| 5 | p.T935M | ACG>ATG | c.2804C>T | Exon 12 |
| 6 | p.N1270S | AAT>AGT | c.3809A>G | Exon 18 |
| 7 | p.R919G | CGG>GGG | c.2755C>G | Exon 12 |
| 8 | p.T888P | ACC>CCC | c.2662A>C | Exon 11 |
| 9 | p.G943D | GGT>GAT | c.2828G>A | Exon 12 |
| 10 | p.S975Y | TCC>TAC | c.2924C>A | Exon 13 |
| 11 | p.E1173K | GAG>AAG | c.3517G>A | Exon 16 |
| 12 | p.D765G | GAC>GGC | c.2294A>G | Exon 8 |
| 13 | p.C490X | TGC>TGA | c.1470C>A | Exon 3 |
| 14 | p.D196E | GAC>GAA | c.588C>A | Exon 2 |
| 15 | p.G250R | GGG>CGG | c.748G>C | Exon 2 |
| 16 | c.1708‐5 T>G | Splicing mutation | c.1708‐5 T>G | Intron 4 |
| 17 | p.I592F | ATC>TTC | c.1774A>T | Exon 5 |
| 18 | p.L692P | CTG>CCG | c.2075 T>C | Exon 6 |
| 19 | p.Y741C | TAT>TGT | c.2222A>G | Exon 8 |
| 20 | p.G869R | GGA>CGA | c.2605G>C | Exon 11 |
| 21 | p.G948C | GGT>GTT | c.2842G>T | Exon 12 |
| 22 | p.G1000R | GGG>CGG | c.2998G>C | Exon 13 |
| 23 | p.I1230T | ATT>ACT | c.3689 T>C | Exon 17 |
| 24 | p.T1286I | ACC>ATC | c.3857C>T | Exon 18 |
Abbreviations: a.a, amino acid; nt, nucleotide.
For the optimization of probes, we tested the DNA microarray for all 24 mutations for WD genotyping using synthesized plasmids as positive samples. The plasmids were amplified by multiplex asymmetric PCR, and the amplified oligonucleotides were identical to those using patient genomic DNA as template. The specificity of the 96 probes was ensured by adjusting their length, the side‐to‐side movement around the mutations, and optimization of hybridization buffer components to reduce the false‐positive and false‐negative results. The probe sequences are listed in Table S7.
3.2. Cut‐off value determination
For determination of the cut‐off value of each probe, a total of 33 clinical samples with known ATP7B variants covering all selected mutations were used. The samples were hybridized to the WD microarray. Eight spots per site can be used for mutation analysis in each matrix, which significantly reduces non‐specific signals. The fluorescence intensity of each site was collected. For genotypes detected in fewer than 10 patients, the fluorescence intensity was recorded by repeating the detection 10 times for statistics and analysis. For cases in which the homozygotes were unavailable, the corresponding plasmids were used for the cut‐off determination. Ratios of the fluorescence intensity of wild‐type probes divided by the mutant probes were calculated. For each loci, the mean value and standard deviation (SD) of the ratios for each genotype were determined by at least ten replicates, and the mean ± 3SD was defined as the cut‐off value for homozygotes, heterozygotes, and wild‐type, respectively (Table 2).
TABLE 2.
Determination of cut‐off value for the 24 loci of ATP7B mutation by the microarray detection.
| No. | Genotype | Homozygotes | Heterozygotes | Wild‐type | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| n | Mean (SD) | CV (%) | Mean ± 3SD | n | Mean (SD) | CV (%) | Mean ± 3SD | n | Mean (SD) | CV (%) | Mean ± 3SD | ||
| 1 | p.D196E | 10 a | 3.11 (0.11) | 3.5 | 2.78–3.44 | 10 | 0.86 (0.02) | 2.3 | 0.80–0.92 | 52 | 0.32 (0.01) | 3.1 | 0.29–0.35 |
| 2 | p.G250R | 10 a | 14.32 (0.50) | 3.5 | 12.82–15.82 | 10 | 0.84 (0.02) | 2.4 | 0.78–0.90 | 52 | 0.17 (0.01) | 5.9 | 0.14–0.20 |
| 3 | p.C490X | 10 a | 3.03 (0.24) | 7.9 | 2.31–3.75 | 10 | 0.95 (0.04) | 4.2 | 0.83–1.07 | 51 | 0.33 (0.02) | 6.1 | 0.27–0.39 |
| 4 | c.1708‐5G>T | 10 a | 3.12 (0.26) | 8.3 | 2.34–3.90 | 10 | 1.06 (0.04) | 3.8 | 0.94–1.18 | 51 | 0.32 (0.01) | 3.1 | 0.29–0.35 |
| 5 | p.I592F | 10 a | 10.13 (0.15) | 1.5 | 9.68–10.58 | 10 | 0.92 (0.08) | 8.7 | 0.68–1.16 | 51 | 0.20 (0.01) | 5 | 0.17–0.23 |
| 6 | p.L692P | 10 a | 8.71 (0.51) | 5.9 | 7.18–10.24 | 10 | 0.85 (0.04) | 4.7 | 0.73–0.97 | 51 | 0.12 (0.01) | 8.3 | 0.09–0.15 |
| 7 | p.Y741C | 10 a | 3.17 (0.20) | 6.3 | 3.11–3.77 | 10 | 0.89 (0.02) | 2.2 | 0.83–0.95 | 52 | 0.32 (0.02) | 6.3 | 0.22–0.38 |
| 8 | p.D765G | 10 | 9.13 (0.28) | 3.1 | 8.29–9.97 | 10 | 1.03 (0.05) | 4.9 | 0.88–1.18 | 51 | 0.21 (0.01) | 4.8 | 0.18–0.24 |
| 9 | p.R778L | 10 | 3.23 (0.08) | 2.5 | 2.99–3.47 | 14 | 1.17 (0.22) | 8.5 | 0.51–1.83 | 36 | 0.31 (0.02) | 6.5 | 0.25–0.37 |
| p.R778Q | 10 a | 3.11 (0.06) | 1.9 | 2.93–3.29 | 1.14 (0.07) | 6.1 | 0.93–1.35 | ||||||
| 10 | p.G869R | 10 a | 9.34 (0.11) | 1.2 | 9.01–9.67 | 10 | 0.86 (0.02) | 2.3 | 0.80–0.92 | 51 | 0.31 (0.01) | 3.2 | 0.28–0.34 |
| 11 | p.A874V | 10 | 4.49 (0.12) | 2.7 | 4.13–4.85 | 10 | 1.27 (0.05) | 3.9 | 1.12–1.42 | 50 | 0.42 (0.03) | 7.1 | 0.33–0.41 |
| 12 | p.T888P | 10 | 2.28 (0.20) | 8.8 | 1.68–2.88 | 10 | 0.84 (0.01) | 1.2 | 0.81–0.87 | 51 | 0.43 (0.02) | 4.7 | 0.37–0.49 |
| 13 | p.R919G | 10 a | 2.87 (0.15) | 5.2 | 2.42–3.32 | 10 | 1.01 (0.01) | 1 | 0.98–1.04 | 52 | 0.34 (0.01) | 2.9 | 0.31–0.37 |
| 14 | p.T935M | 10 a | 2.24 (0.17) | 7.6 | 1.70–2.78 | 10 | 1.65 (0.02) | 3.3 | 1.59–1.71 | 51 | 0.63 (0.05) | 7.9 | 0.48–0.78 |
| 15 | p.G943D | 10 a | 10.07 (0.04) | 0.4 | 9.95–10.19 | 10 | 1.79 (0.04) | 2.2 | 1.67–1.91 | 50 | 0.70 (0.01) | 1.4 | 0.67–0.73 |
| 16 | p.G948C | 10 a | 4.25 (0.35) | 8.2 | 3.20–5.30 | 10 | 1.06 (0.02) | 1.9 | 1.00–1.12 | 52 | 0.74 (0.03) | 4.1 | 0.65–0.83 |
| 17 | p.S975Y | 10 a | 6.28 (0.22) | 3.5 | 5.62–6.94 | 10 | 0.92 (0.07) | 7.6 | 0.71–1.13 | 49 | 0.36 (0.01) | 2.8 | 0.33–0.39 |
| 18 | p.G1000R | 10 | 12.94 (0.08) | 0.6 | 12.7–13.18 | 10 | 1.50 (0.03) | 2 | 1.41–1.59 | 51 | 0.38 (0.01) | 2.6 | 0.35–0.41 |
| 19 | p.V1106I | 10 | 3.26 (0.27) | 8.3 | 2.45–4.07 | 10 | 0.85 (0.02) | 2.3 | 0.79–0.91 | 49 | 0.31 (0.01) | 3.2 | 0.28–0.34 |
| 20 | p.I1148T | 10 | 3.20 (0.16) | 5 | 2.72–3.68 | 10 | 0.81 (0.03) | 3.7 | 0.72–0.90 | 51 | 0.31 (0.01) | 3.2 | 0.28–0.34 |
| 21 | p.E1173K | 10 a | 4.99 (0.10) | 2 | 4.69–5.29 | 10 | 0.88 (0.02) | 2.2 | 0.82–0.94 | 52 | 0.20 (0.01) | 5 | 0.17–0.23 |
| 22 | p.I1230T | 10 | 16.97 (0.14) | 0.8 | 16.55–17.39 | 10 | 1.29 (0.03) | 2.3 | 1.20–1.38 | 51 | 0.46 (0.01) | 2.2 | 0.43–0.49 |
| 23 | p.N1270S | 10 a | 3.72 (0.27) | 7.3 | 2.91–4.53 | 10 | 1.03 (0.04) | 3.9 | 0.91–1.15 | 51 | 0.27 (0.01) | 3.7 | 0.24–0.30 |
| 24 | p.T1286I | 10 a | 7.13 (0.48) | 6.7 | 5.69–8.57 | 10 | 1.02 (0.04) | 3.9 | 0.90–1.14 | 51 | 0.54 (0.02) | 3.7 | 0.48–0.60 |
Plasmids were used for the cases without homozygous mutation. Mean = average of ratio, ratio = signal value of wild‐type probes divided by signal value of mutant probes.
3.3. Microarray sensitivity, specificity and repeatability
Genomic DNA from 10 patients was subjected to gradient dilution to obtain DNA solutions of 5, 10, 20 and 50 ng/μl, with a total amount of 15, 30, 60 and 150 ng DNA (The amount added to each tube in multiplex asymmetric PCR). The probe hybridization signal represented by each DNA concentration was the mean of all positive probes in the 10 samples involved in the hybridization. The results indicated that the signal value of the microarray hybridization probes was high when the DNA concentration was 20–50 ng/μl, and the signal was low when the concentration was lower than 10 ng/μl. A 10 ng/μl (30 ng/test) wild‐type sample result is shown in Figure 2A,B. The test was considered as failed when the genomic DNA was 5 ng/μl (15 ng/test). Thus, the sensitivity (minimum detection limit) is 30 ng/test. The results detected by the microarray showed the coincidence rate of 100% with that detected by Sanger sequencing (Table 2).
FIGURE 2.
Representative result of multiplex asymmetric PCR and microarray detection. (A) Twelve bands of PCR products covering all detection sites are distributed in three tubes. The length of the bands and the target fragments in the three tubes are as follows: Tube 1, 287 bp (p.Y741C, p.D765G, p.R778L/Q), 250 bp (p.N1270S, p.T1286I), 172 bp (p.L692P), 137 bp (p.I1148T, p.E1173K); tube 2, 248 bp (p.R919G, p.T935M, p.G943D, p.G948C), 212 bp (p.G869R, p.A874V, p.T888P), 166 bp (p.S975Y, p.G1000R), 130 bp (p.V1106I); and tube 3, 260 bp (p.D196E, p.G250R), 226 bp (p.C490X), 196 bp (c.1708‐5 T>G, p.I592F), and 84 bp (p.I1230T). (B) A representative genotyping result of a clinical sample by microarray detection.
The reproducibility of the microarray was also investigated by calculating the coefficient of variation (CV) value of the hybridization signal of each probe, and the fluorescence ratio was calculated with the 33 cases with the 24 different known ATP7B variants. Different matrices of the same microarray were applied to test the same samples. All samples were tested 10 times to calculate the CV value (Table 2). The CV values of the fluorescence ratio of all mutations were less than 10%, confirming good reproducibility of the WD microarray.
3.4. Clinical evaluation of WD microarray accuracy
We assessed our WD microarray using DNA from 126 samples, including 106 suspected WD samples and 20 patients with other liver diseases as a negative control for clinical evaluation. The microarray results revealed that 86 of the 106 (81.13%) suspected WD cases harbored ATP7B mutations. Sanger sequencing of the other 20 cases identified other mutations in ATP7B not included in the microarray. No ATP7B mutation was identified in patients with other liver diseases. For the 24 mutations included in the microarray, the coincidence rate was 100% between the results detected by the microarray and the sequencing results (Table 3). We also calculated the mutation frequency of the 106 WD cases. The results showed that the mutation frequency of p.R778L (50.94%), p.A874V (17.92%), p.P992L (11.32%), p.V1106I (11.32%), and p.I1148T (6.60%) were over 5% among all 24 mutations. The frequency of the identified mutations is shown in Figure 3. In addition, 20 mutations not included in the microarray were identified by sequencing (Table 3), including c.3659‐3660insTGA, which was reported for the first time.
TABLE 3.
Results of microarray detection of ATP7B mutation in 106 suspected WD cases and comparison with that by sequencing.
| No. | Cases No. | WD microarray | Sanger sequencing | Coincidence of tested mutations | ||
|---|---|---|---|---|---|---|
| Mutation | Ratio of signal values | Positive/Negative | ||||
| 1 | SWD1 |
p.G250R p.R778L NA |
0.83 1.22 NA |
Positive Positive NA |
p.G250R p.R778L p.R616Q |
Yes |
| 2 | SWD2 | NA | NA | NA | p.G881D | Yes |
| 3 | SWD3 | p.R778L | 0.62 | Positive | p.R778L | Yes |
| 4 | SWD4 | p.R778L | 1.59 | Positive | p.R778L | Yes |
| 5 | SWD5 |
p.Y741C p.N1270S |
0.93 1.07 |
Positive Positive |
p.Y741C p.N1270S |
Yes |
| 6 | SWD6 | p.R778L | 1.68 | Positive | p.R778L | Yes |
| 7 | SWD7 | p.R778L | 1.04 | Positive | p.R778L | Yes |
| 8 | SWD8 |
p.R778L p.V1106I |
1.44 0.91 |
Positive Positive |
p.R778L p.V1106I |
Yes |
| 9 | SWD9 | p.A874V | 1.26 | Positive | p.A874V | Yes |
| 10 | SWD10 |
p.A874V p.V1106I |
1.28 0.80 |
Positive Positive |
p.A874V p.V1106I |
Yes |
| 11 | SWD11 | p.R778L | 1.50 | Positive | p.R778L | Yes |
| 12 | SWD12 |
p.T935M p.I1148T |
1.63 0.87 |
Positive Positive |
p.T935M p.I1148T |
Yes |
| 13 | SWD13 | p.V1106I | 0.85 | Positive | p.V1106I | Yes |
| 14 | SWD14 | p.I1148T | 0.79 | Positive | p.I1148T | Yes |
| 15 | SWD15 |
p.R778L p.E1173K |
1.37 0.86 |
Positive Positive |
p.R778L p.E1173K |
Yes |
| 16 | SWD16 |
NA NA |
NA NA |
NA NA |
p.V1297I p.G205V |
Yes |
| 17 | SWD17 | p.N1270S | 1.07 | Positive | p.N1270S | Yes |
| 18 | SWD18 |
p.R778L (Homo) p.V1106I |
3.22 0.84 |
Positive Positive |
p.R778L (Homo) p.V1106I |
Yes |
| 19 | SWD19 | p.R778L (Homo) | 3.06 | Positive | p.R778L (Homo) | Yes |
| 20 | SWD20 | p.G943D | 1.73 | Positive | p.G943D | Yes |
| 21 | SWD21 | p.I1148T | 0.75 | Positive | p.I1148T | Yes |
| 22 | SWD22 |
p.R778L p.I1148T |
1.16 0.73 |
Positive Positive |
p.R778L p.I1148T |
Yes |
| 23 | SWD23 |
p.R778L (Homo) p.V1106I |
3.47 0.85 |
Positive Positive |
p.R778L (Homo) p.V1106I |
Yes |
| 24 | SWD24 | p.R778L (Homo) | 3.24 | Positive | p.R778L (Homo) | Yes |
| 25 | SWD25 | p.I1148T | 0.89 | Positive | p.I1148T | Yes |
| 26 | SWD26 |
p.R778L p.G943D |
1.12 1.81 |
Positive Positive |
p.R778L p.G943D |
Yes |
| 27 | SWD27 | p.A874V | 1.13 | Positive | p.A874V | Yes |
| 28 | SWD28 |
NA NA |
NA NA |
NA NA |
p.P663R p.L1188F |
Yes |
| 29 | SWD29 | p.R778L | 1.19 | Positive | p.R778L | Yes |
| 30 | SWD30 | p.R778L | 1.37 | Positive | p.R778L | Yes |
| 31 | SWD31 | p.R778L | 1.25 | Positive | p.R778L | Yes |
| 32 | SWD32 | p.R778L (Homo) | 3.11 | Positive | p.R778L (Homo) | Yes |
| 33 | SWD33 |
p.R778L NA |
1.37 NA |
Positive NA |
p.R778L p.P992L |
Yes |
| 34 | SWD34 | NA | NA | NA | K1010T | Yes |
| 35 | SWD35 | p.A874V | 1.33 | Positive | p.A874V | Yes |
| 36 | SWD36 | p.R778L | 1.61 | Positive | p.R778L | Yes |
| 37 | SWD37 |
NA NA |
NA NA |
NA NA |
p.V997L p.A1168T |
Yes |
| 38 | SWD38 | p.R778L | 1.27 | Positive | p.R778L | Yes |
| 39 | SWD39 | p.R778L | 1.31 | Positive | p.R778L | Yes |
| 40 | SWD40 | p.R778L | 1.29 | Positive | p.R778L | Yes |
| 41 | SWD41 |
p.R778L p.V1106I |
1.26 0.85 |
Positive Positive |
p.R778L p.V1106I |
Yes |
| 42 | SWD42 | p.I1148T | 0.76 | Positive | p.I1148T | Yes |
| 43 | SWD43 | NA | NA | NA | p.P992L | Yes |
| 44 | SWD44 | p.R778L | 1.31 | Positive | p.R778L | Yes |
| 45 | SWD45 | p.R778L | 1.34 | Positive | p.R778L | Yes |
| 46 | SWD46 | p.R778L | 1.25 | Positive | p.A874V | Yes |
| 47 | SWD47 |
p.A874V NA |
1.35 NA |
Positive NA |
p.A874V p.P992L (Homo) |
Yes |
| 48 | SWD48 | p.R778L | 1.35 | Positive | p.R778L | Yes |
| 49 | SWD49 | p.R778L | 1.34 | Positive | p.R778L | Yes |
| 50 | SWD50 | p.A874V | 1.32 | Positive | p.A874V | Yes |
| 51 | SWD51 | p.A874V | 1.39 | Positive | p.A874V | Yes |
| 52 | SWD52 | p.R778L | 1.33 | Positive | p.R778L | Yes |
| 53 | SWD53 | NA | NA | NA | p.P992L | Yes |
| 54 | SWD54 | p.A874V | 1.29 | Positive | p.A874V | Yes |
| 55 | SWD55 | NA | NA | NA | p.P992L | Yes |
| 56 | SWD56 | p.V1106I | 0.83 | Positive | p.V1106I | Yes |
| 57 | SWD57 | NA | NA | NA | p.P992L | Yes |
| 58 | SWD58 |
p.R778L p.V1106I |
1.32 0.80 |
Positive Positive |
p.R778L p.V1106I |
Yes |
| 59 | SWD59 | p.R778L | 1.22 | Positive | p.R778L | Yes |
| 60 | SWD60 | p.R778L | 1.26 | Positive | p.R778L | Yes |
| 61 | SWD61 | p.R778L | 1.81 | Positive | p.R778L | Yes |
| 62 | SWD62 |
p.D196E p.V1106I (Homo) NA |
2.87 0.83 NA |
Positive Positive NA |
p.D196E p.V1106I (Homo) c.3903+6C>T |
Yes |
| 63 | SWD63 |
p.R778L NA NA |
1.65 NA NA |
Positive NA NA |
p.R778L c.3903+6C>T p.R1151H |
Yes |
| 64 | SWD64 | NA | NA | NA | c.3903+6C>T | Yes |
| 65 | SWD65 | NA | NA | NA | c.3903+6C>T | Yes |
| 66 | SWD66 |
p.V1106I NA NA NA NA |
0.86 NA NA NA NA |
Positive NA NA NA NA |
p.V1106I c.3903+6C>T p.A887L p.A874P p.C154S |
Yes |
| 67 | SWD67 |
NA NA |
NA NA |
NA NA |
p.V1297I p.T766M |
Yes |
| 68 | SWD68 | p.R778L | 1.18 | Positive | p.R778L | Yes |
| 69 | SWD69 | p.R778L (Homo) | 3.46 | Positive | p.R778L (Homo) | Yes |
| 70 | SWD70 |
p.G943D NA |
1.80 NA |
Positive NA |
p.G943D p.P992L |
Yes |
| 71 | SWD71 | p.R778L | 1.66 | Positive | p.R778L | Yes |
| 72 | SWD72 | NA | NA | NA | p.V1297I | Yes |
| 73 | SWD73 |
p.A874V NA |
1.18 NA |
Positive NA |
p.A874V p.V606G |
Yes |
| 74 | SWD74 |
p.A874V NA |
1.34 NA |
Positive NA |
p.A874V c.3659‐3660ins TGA |
Yes |
| 75 | SWD75 |
NA NA |
NA NA |
Positive NA |
p.P992L (Homo) p.R969W |
Yes |
| 76 | SWD76 | p.R778L | 0.70 | Positive | p.R778L | Yes |
| 77 | SWD77 | p.R778L | 1.74 | Positive | p.R778L | Yes |
| 78 | SWD78 | p.A874V | 1.30 | Positive | p.A874V | Yes |
| 79 | SWD79 | p.A874V | 1.40 | Positive | p.A874V | Yes |
| 80 | SWD80 | p.V1106I | 0.87 | Positive | p.V1106I | Yes |
| 81 | SWD81 |
NA NA |
NA NA |
NA NA |
c.3903+6C>T p.V1216M |
Yes |
| 82 | SWD82 |
p.R778L p.A874V |
1.39 1.35 |
Positive Positive |
p.R778L p.A874V |
Yes |
| 83 | SWD83 | p.R778L | 0.55 | Positive | p.R778L | Yes |
| 84 | SWD84 | p.A874V | 1.33 | Positive | p.A874V | Yes |
| 85 | SWD85 | NA | NA | NA | p.P992L | Yes |
| 86 | SWD86 | p.R778L (Homo) | 3.19 | Positive | p.R778L (Homo) | Yes |
| 87 | SWD87 | p.R778L | 0.96 | Positive | p.R778L | Yes |
| 88 | SWD88 |
NA NA |
NA NA |
NA NA |
p.W779X p.A874P |
Yes |
| 89 | SWD89 |
p.R778L p.A874V |
1.07 1.13 |
Positive Positive |
p.R778L p.A874V |
Yes |
| 90 | SWD90 |
p.R778L p.A874V |
1.61 1.19 |
Positive Positive |
p.R778L p.A874V |
Yes |
| 91 | SWD91 | p.V1106I | 0.89 | Positive | p.V1106I | Yes |
| 92 | SWD92 | NA | NA | NA | p.P992L | Yes |
| 93 | SWD93 | p.A874V | 1.32 | Positive | p.A874V | Yes |
| 94 | SWD94 | p.R778L | 0.74 | Positive | p.R778L | Yes |
| 95 | SWD95 | p.R778L | 1.58 | Positive | p.R778L | Yes |
| 96 | SWD96 | p.R778L | 0.58 | Positive | p.R778L | Yes |
| 97 | SWD97 | p.I1148T | 0.77 | Positive | p.I1148T | Yes |
| 98 | SWD98 | NA | NA | NA | p.P992L | Yes |
| 99 | SWD99 | p.R778L | 1.06 | Positive | p.R778L | Yes |
| 100 | SWD100 | p.R778L | 1.35 | Positive | p.R778L | Yes |
| 101 | SWD101 | p.A874V | 1.27 | Positive | p.A874V | Yes |
| 102 | SWD102 | p.S975Y | 0.98 | Positive | p.S975Y | Yes |
| 103 | SWD103 | p.R778L | 0.88 | Positive | p.R778L | Yes |
| 104 | SWD104 | NA | NA | NA | p.P992L | Yes |
| 105 | SWD105 | p.R778L | 1.74 | Positive | p.R778L | Yes |
| 106 | SWD106 | p.R778L | 0.66 | Positive | p.R778L | Yes |
Abbreviations: Homo, Homozygote; NA, not available as the loci were not included in the microarray.
FIGURE 3.
Schematic representation of the prevalence of individual ATP7B mutations in Chinese WD patients as detected by the microarray method. p.R778L (50.94%), p.A874V (17.92%), p.P992L (11.32%), p.V1106I (11.32%), and p.I1148T (6.60%) were the most common mutations. ATP7B p.P992L was identified by sequencing.
Among the 106 WD cases, 7, 2 and 1 cases had a homozygous mutation of p.R778L, p.R992L, and p.V1106I, respectively, and the other cases harbored heterozygous mutations (Table 3).
We also analyzed the correlation between the representative clinical presentations such as gender, age, symptoms at onset, ceruloplasmin level of the 106 suspected WD cases (Table S2) and the prevalence of the top five hotspot mutations (p.R778L, p.A874V, p.P992L, p.V1106I, and p.I1148T). The results showed that there was no significant correlation between the clinical parameters with the individual mutations (Figure S1).
4. DISCUSSION
In the present study, we developed and analyzed a microarray method, with advantages of simplicity, efficiency, cost‐effectiveness, and user‐friendliness, for screening hotspot ATP7B variants in Chinese WD patients. Twenty‐four of the most ATP7B mutations, covering up to 70% of Chinese patients with WD reported to date, were simultaneously detected by the microarray. This microarray is a reliable detection system with an approximate 6 h testing time. Clinical evaluation of the microarray showed that accuracy for the detection of ATP7B mutations reached 100%.
Different from the mutations detected in WD cases in Asia, p.H1069Q (c.3207C>A) is the most common mutation in WD patients from European countries. In Asia, p.R778L (c.2333G>T, exon 8) is the most common ATP7B mutation, with a frequency of 10.52% in Thailand, 11 similar to that in other Asian countries. In a recent study of WD cases in South China, ATP7B mutation mainly occurred in exon 8 (23.30%) and exon 16 (12.14%), and the top three mutations were p.R778L (18.93%), p.I1148T (8.74%), and p.P992L (4.37%). 8 This result is consistent with a previous study in eastern China, in which the top three mutations in WD cases were p.R778L (31.9%), p.P992L (11.2%), and p.A874V (5.17%). 9 In Hong Kong and Taiwan, p.R778L accounted for 17.3% 6 and 29.63% 12 of the reported ATP7B mutations, respectively. A systematic literature review of 345 Chinese cases showed that the most common ATP7B mutations were located in exons 8, 13, 12 and 16, accounting for 74.0% of the reported ATP7B mutations and the most common ATP7B mutations were p.R778L and p.P992L, accounting for 50.43% of all reported ATP7B alleles. Similarly, in Korea, the allele frequency of p.R778L was 39.2%. 13 This is in sharp contrast to a study in Japan that revealed that the most common of the 13 mutations in 23 Japanese WD cases were c.2874delC (exon 13, 30%) and p.R778L (exon 8, 25%). 11
Our microarray results showed that the mutation frequency of p.R778L (50.94%), p.A874V (17.92%), p.V1106I (11.32%), and p.I1148T (6.60%) were over 5% in Chinese patients with WD. In addition, p.G943D (2.83%), p.T888P (1.89%), and p.N1270S (1.89%) showed relatively high mutation frequency. p.D196E, p.G250R, p.Y741C, p.T935M, p.S975Y, and p.E1173K were each identified in one sample. No samples contained any of the remaining 11 mutations. These results were consistent with the previous literature and related mutation database reported in Chinese patients with WD. In addition, Sanger sequencing showed that the most frequent mutation, outside of the 24 mutations examined, was p.R992L (11.32%), followed by the splicing mutation c.3903+6C>T (5.56%), p.V1297I (2.83%), and p.A874P(1.89%). Our results identified 21 mutations that were each present in one sample, which may further improve the mutation database (Figure 3). p.V1106I was one of the most common ATP7B mutation reported by our previous study. 5 However, our results need to be confirmed in a larger cohort of WD cases. We also analyzed the correlation between the clinical parameters such as gender, age, symptom at onset and ceruloplasmin level of the WD cases and the prevalence of the top five hotspot mutations (p.R778L, p.A874V, p.P992L, p.V1106I, and p.I1148T). Our results showed that there was no significant correlation between the clinical parameters and the individual mutation, possibly due to the small sample size.
Early diagnosis of WD has a major impact on the prognosis of the disease. Misdiagnosis can cause severe liver failure or even death. WD diagnosis is currently based on a range of clinical features, including Kayser‐Fleischer rings, abnormal liver function, hepatic failure, and neurological symptoms. To date, Sanger sequencing remains the most widely used method for detecting potential ATP7B genetic variants. However, with 21 exons in the ATP7B gene, Sanger sequencing is not convenient for independent sample analysis. NGS that enables high‐throughput and full‐sequence processing has been used in research and clinical practice. 14 However, this approach has a long running time, high cost, a complicated procedure and data pruning and analysis. The time periods for analysis are generally three to 5 days, or longer, from sampling to sequencing. WD has relatively concentrated pathogenic mutations, and thus it is wasteful to use NGS to detect the entire genome or exome. Furthermore, traditional genetic testing methods are time‐consuming and costly, and only several loci can be detected in one test; therefore, these methods are not suitable for rapid clinical genetic screening. A simple, low‐cost, efficient, and highly targeted detection method is required for WD genetic testing to better assist clinical diagnosis and treatment.
As there are common ethnic‐specific mutations in ATP7B among Chinese WD patients, the microarray is an efficient, cost‐effective and convenient method for simultaneous detection of a series of ATP7B variants in multiple independent DNA samples. 14 A previous study reported a microarray for the genotyping of ATP7B variants in Czech and Slovak populations using an arrayed primer extension (APEX) reaction method. However, this approach requires 24 h, leading to complicated operation and difficulties in primer set design and microarray preparation, all of which increase the direct cost. 15 Moreover, this technology has high DNA quality requirements, which increases the associated time and cost because the DNA needs to be extensively purified. Manjula and co‐workers reported a microarray that was based on the amplified PCR products that labelled with digoxigenin. 16 The sample preparing was complicated and time‐consuming because of the nested PCR and the following digoxigenin labelling. Moreover, the designed probes only focused on the wild‐type and most common mutated bases, the precise genotype was impossible. Finally, and most importantly, the use of colorimetric detection and a simple home‐made flat‐bed scanner was not as sensitive as the fluorescence based methods. 16 In our WD microarray, the Cy3 fluorophore is directly added to the 5′ end of the reverse primer; Cy3‐labeled single‐stranded DNA fragments can be easily harvested using multiplex asymmetric PCR, and the fluorescence intensity is captured by a microarray scanner. Our WD microarray takes approximately 6 hrs from DNA extraction to identification of the 24 most common ATP7B mutations. This approach is also highly automated, standardized, and easy to operate, making it suitable for clinical diagnosis. Indeed, for future commercial applications in clinics, the microarray needs to be validated by other laboratories. We will conduct the validation of the microarray in other laboratories in the near future.
Notably, this microarray method is highly dependent on existing understanding of WD. Furthermore, we failed to detect any mutations in several patients included in our study group. Both false‐positive and false‐negative results cannot be avoided if patients harbor some specific rare mutations. For example, because of the short distance from the hotspot sites, the p.W779X variant may reduce the accuracy of the probes for the p.R778L variant, resulting in aberrant fluorescence intensity. Moreover, the p.P992L variant is also a hotspot in Chinese WD patients, but the GC content in the sequences near the mutation site is extremely high. This makes it difficult to balance probe hybridization temperature, and the hybridization temperature required for a probe at this site is much higher than other probes. Therefore, the p.P992L variant was not included in our WD microarray. However, based on the high‐resolution melting curve (HRM), we have developed a detection method for p.P992L and applied for a patent in order to make up for the shortcomings of this site. 17 With the identification of novel recurrent ATP7B mutation and specifically, hotspot mutations in other countries such as p.H1069Q in the western countries, more probes can be included in the microarray, and this method can be used for the detection of more ATP7B mutations in patients worldwide.
In summary, our WD microarray represents a potential convenient tool for the clinical screening of ATP7B gene mutations, especially in Chinese WD patients. This method will be helpful in the diagnosis, genetic counseling and treatment of WD patient.
AUTHOR CONTRIBUTIONS
All authors accept responsibility for the entire content of this manuscript and approved its submission. JH and DHZ designed the study. SYJ and DHZ analyzed and interpreted the data and wrote the manuscript. SYJ, DHZ, XJL, and BZ performed the experiments. WZ, WJD and ZW provided patient samples and clinical data. XJO, JH, and DHZ advised on the conception of the study and supervised the study. JH and DHZ revised the paper. All authors vouch for the data and analysis, approved the final version, and agree to publish the manuscript.
FUNDING INFORMATION
This study was supported by grants from the Digestive Medical Coordinated Development Center of Beijing Municipal Administration of Hospitals (No. XXX0101, XXZ0501, and XXT03) and the Beijing Natural Science Foundation (No. 7222036).
CONFLICT OF INTEREST
The authors have no conflict of interest related to this study.
Supporting information
Figure S1
Table S1
Table S2
Table S3
Table S4
Table S5
Table S6
Table S7
ACKNOWLEDGEMENT
We thank Gabrielle White Wolf, PhD, from Liwen Bianji, Edanz Group China (https://www.liwenbianji.cn/), for editing the English text of a draft of this manuscript. The abstract of present study had been accepted for oral presentation by the Committee of the 29th Asian Pacific Association for the Study of the Liver (APASL) in 2020.
Jia S, Li X, Zhang W, et al. Laboratory and clinical evaluation of a microarray for the detection of ATP7B mutations in Wilson disease in China. J Clin Lab Anal. 2022;36:e24735. doi: 10.1002/jcla.24735
Contributor Information
Donghu Zhou, Email: donghuzhou@gmail.com.
Jian Huang, Email: huangj1966@hotmail.com.
DATA AVAILABILITY STATEMENT
The data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.
REFERENCES
- 1. Bandmann O, Weiss KH, Kaler SG. Wilson's disease and other neurological copper disorders. Lancet Neurol. 2015;14(1):103‐113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Dziezyc K, Litwin T, Chabik G, Gramza K, Czlonkowska A. Families with Wilson's disease in subsequent generations: clinical and genetic analysis. Mov Disord. 2014;29(14):1828‐1832. [DOI] [PubMed] [Google Scholar]
- 3. Ferenci P, Caca K, Loudianos G, et al. Diagnosis and phenotypic classification of Wilson disease. Liver Int. 2003;23(3):139‐142. [DOI] [PubMed] [Google Scholar]
- 4. Czlonkowska A, Litwin T, Dusek P, et al. Wilson disease. Nat Rev Dis Primers. 2018;4(1):21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Li X, Zhang W, Zhou D, et al. Complex ATP7B mutation patterns in Wilson disease and evaluation of a yeast model for functional analysis of variants. Hum Mutat. 2019;40(5):552‐565. [DOI] [PubMed] [Google Scholar]
- 6. Mak CM, Lam CW, Tam S, et al. Mutational analysis of 65 Wilson disease patients in Hong Kong Chinese: identification of 17 novel mutations and its genetic heterogeneity. J Hum Genet. 2008;53(1):55‐63. [DOI] [PubMed] [Google Scholar]
- 7. Wan L, Tsai CH, Hsu CM, et al. Mutation analysis and characterization of alternative splice variants of the Wilson disease gene ATP7B. Hepatology. 2010;52(5):1662‐1670. [DOI] [PubMed] [Google Scholar]
- 8. Wei Z, Huang Y, Liu A, et al. Mutational characterization of ATP7B gene in 103 Wilson's disease patients from Southern China: identification of three novel mutations. Neuroreport. 2014;25(14):1075‐1080. [DOI] [PubMed] [Google Scholar]
- 9. Dong Y, Ni W, Chen WJ, et al. Spectrum and classification of ATP7B Variants in a large cohort of Chinese patients with Wilson's Disease guides genetic diagnosis. Theranostics. 2016;6(5):638‐649. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Hua R, Hua F, Jiao Y, et al. Mutational analysis of ATP7B in Chinese Wilson disease patients. Am J Transl Res. 2016;8(6):2851‐2861. [PMC free article] [PubMed] [Google Scholar]
- 11. Panichareon B, Taweechue K, Thongnoppakhun W, et al. Six novel ATP7B mutations in Thai patients with Wilson disease. Eur J Med Genet. 2011;54(2):103‐107. [DOI] [PubMed] [Google Scholar]
- 12. Lee CC, Wu JY, Tsai FJ, et al. Molecular analysis of Wilson disease in Taiwan: identification of one novel mutation and evidence of haplotype‐mutation association. J Hum Genet. 2000;45(5):275‐279. [DOI] [PubMed] [Google Scholar]
- 13. Park S, Park JY, Kim GH, et al. Identification of novel ATP7B gene mutations and their functional roles in Korean patients with Wilson disease. Hum Mutat. 2007;28(11):1108‐1113. [DOI] [PubMed] [Google Scholar]
- 14. Poujois A, Woimant F. Challenges in the diagnosis of Wilson disease. Ann Transl Med. 2019;7(Suppl 2):S67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Gojova L, Jansova E, Kulm M, Pouchla S, Kozak L. Genotyping microarray as a novel approach for the detection of ATP7B gene mutations in patients with Wilson disease. Clin Genet. 2008;73(5):441‐452. [DOI] [PubMed] [Google Scholar]
- 16. Mathur M, Singh E, Poduval TB, Rao AV. Development of low‐density oligonucleotide microarrays for detecting mutations causing Wilson's disease. Indian J Med Res. 2015;141(2):175‐186. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Xu A, Lv T, Zhang B, Zhang W, Ou X, Huang J. Development and evaluation of an unlabeled probe high‐resolution melting assay for detection of ATP7B mutations in Wilson's disease. J Clin Lab Anal. 2017;31(4):e22064. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1
Table S1
Table S2
Table S3
Table S4
Table S5
Table S6
Table S7
Data Availability Statement
The data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.