| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 ARUP Institute for Clinical and Experimental Pathology, ARUP Laboratories, 500 Chipeta Way, Salt Lake City, UT 84108.
2 Department of Pathology, University of Utah School of Medicine, Salt Lake City, UT 84132.
aAuthor for correspondence. Fax 801-584-5114; e-mail eralimc{at}aruplab.com.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: The NanoChip® Molecular Biology Workstation (Nanogen) uses electronic microarrays for mutation detection. Factor V, factor II, and MTHFR genotypes identified in the NanoChip system on 225 samples were compared with genotypes from LightCycler® assays (Roche). We determined within- and between-cartridge signal and ratio variation and analyzed the effect of additional mutations at or near the detection area used for the NanoChip assays.
Results: Genotypes determined for all three mutations on the NanoChip platform were in complete concordance with LightCycler results. Within-cartridge signal variation as measured by the CV of fluorescence signals was <10% for each allele when present. The within-cartridge CV for heterozygous mutant/wild-type ratios was <8.5%, and between-cartridge CV was <18%. A dilution study showed that results could be obtained in this assay with 6 ng of nucleic acid per PCR, the lowest input tested. The presence of additional sequence variations near the expected mutations can produce equivocal or discrepant results.
Conclusions: Mutation detection using the NanoChip Molecular Biology Workstation was accurate and reproducible for the three assays evaluated.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Molecular methods such as allele-specific amplification, PCR-restriction fragment length polymorphism analysis, allele-specific oligonucleotide hybridization, non-PCR oligonucleotide cleavage technology, and real-time PCR have all been applied to the detection of mutations associated with thrombophilia (5)(6)(7)(8)(9). This study presents an evaluation of an electronic microarray-based DNA analysis for the presence of these mutations (10).
We used the NanoChip® Molecular Biology Workstation and the microchip cartridge that contains 10 x 10 electronically addressable sites (Nanogen). The assays for the factor V 1691 mutation and factor II 20210 mutation use biotinylated amplicons from a multiplex PCR with a separate PCR for the MTHFR 677 mutation. Amplicons were loaded onto chips by electronic activation of specific locations. Streptavidin in the gel covering the microarray facilitated immobilization of amplicons. Hybridization of fluorescently labeled reporter probes was made specific by choice of incubation temperatures. The fluorescence signals from mutant and wild-type probes were analyzed as fluorescence ratios to determine genotype. We compared the results to genotypes determined by LightCyclerTM assays. We examined the variation of the NanoChip system and the minimum template required. The effect of additional sequence variations near the expected mutations for factor II and MTHFR was also evaluated.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
NanoChip ASSAYS
PCR amplification and desalting.
Oligonucleotides, primers, and probes used in the NanoChip assays were provided by Nanogen (San Diego, CA). The sequences are listed in Table 1
. The factor V and factor II NanoChip assays were performed according to instructions in the Nanogen Factor V (Leiden)/Prothrombin Research Application Note (13) with reagents and instrumentation supplied by Nanogen.
|
The PCR for MTHFR was performed according to instructions in the Nanogen MTHFR Application Note (14). The multiplex PCR for factor V and factor II was performed in a reaction volume of 25 µL with 0.25 µM each of the factor V primers, 0.20 µM each of the factor II primers, 0.2 mM each of the deoxynucleotide triphosphates, 2.5 mM MgCl2, 50 mM KCl, 10 mM Tris-HCl (pH 8.3), and 0.65 U of AmpliTaq Gold. The MTHFR PCR was the same as for the factor V/factor II multiplex assay with primer concentrations for MTHFR of 0.20 µM. A volume of 2 µL of genomic DNA was added to each PCR. Thermocycling was done on a 96-well GeneAmp PCR System 9700 (Applied Biosystems) with the following settings: 1 cycle at 95 °C for 5 min; 40 cycles with denaturation at 94 °C for 20 s, annealing at 55 °C for 20 s, and elongation at 72 °C for 45 s; 1 cycle at 72 °C for 5 min; and a final hold at 4 °C.
The multiplex PCR for factor V and factor II yielded 220- and 163-bp products, respectively, and the MTHFR PCR produced a 173-bp amplicon. Amplicons were desalted with the MultiScreen® PCR 96-well Filtration System (Millipore). For desalting, PCR product was added to the 96-well plate, and vacuum was applied at 250 mmHg negative pressure. The filtration plate was washed once with 100 µL of water per well, and 150 µL of 50 mmol/L L-histidine (conductivity <100 µS/cm) was added to each well. The plate was placed on a Titer Plate Shaker (LabLine Instruments) set at speed 5 for 5 min. The 150 µL of histidine buffer containing the amplicon was pipetted vigorously across the membrane at least 10 times to ensure optimal recovery and then transferred to a clean PCR plate for storage.
Heterozygous ratio reference controls.
Biotinylated mutant and wild-type synthetic oligonucleotides that would bind to either Cy5 or Cy3 probes served as heterozygous ratio reference controls that were used to normalize the Cy5 and Cy3 (mutant and wild-type) fluorescence signals of the clinical samples. The heterozygous ratio reference controls were prepared by mixing equal concentrations of the biotinylated mutant and wild-type oligonucleotides. Various concentrations of oligonucleotides were included on each run to ensure adequate signal coverage, usually 4–6 nmol/L.
Cartridge loading.
Sample amplicons, control amplicons, and heterozygous ratio references were automatically loaded on specific array sites on the 100-site microarray NanoChip cartridge by microelectronic addressing; loading was performed in the NanoChip Loader. Attachment to specific sites was facilitated by streptavidin binding of biotinylated amplicons and controls (15). The double-stranded amplicons were denatured with 0.1 mol/L NaOH for 3 min; the cartridge was then manually washed with a high-salt buffer containing 50 mmol/L sodium phosphate, 500 mmol/L sodium chloride (pH 7.4). Loaded cartridges could be stored at 4 °C when covered with water to avoid crystal formation.
Reporter hybridization.
A reporter/stabilizer mixture for each assay was prepared that contained 0.5 µmol/L mutant reporter probe labeled with Cy5, 0.5 µmol/L wild-type reporter probe labeled with Cy3, and 1 µmol/L stabilizer in high-salt buffer. The cartridge microarray was covered with the reporter/stabilizer mixture for 3 min at room temperature to allow reporter probes and stabilizers to hybridize. For the factor V/factor II multiplex assay, the cartridge microarray was first probed with the factor V reporter probes and detected. The factor V reporter probes were then removed with 0.1 mol/L NaOH, and the factor II reporter/stabilizer mixture was added to the cartridge microarray.
Detection.
After reporter hybridization, the NanoChip cartridge was placed in the NanoChip Reader, and excess reporter probe and stabilizer were removed with high-salt buffer washes. The temperature of the microchip environment was increased to a discrimination temperature at which mismatched reporter probes were denatured and matched reporter probes remained hybridized. The discrimination temperatures were 32, 37, and 40 °C for factor V, factor II, and MTHFR, respectively. Unbound probes were removed with four washes with a low-salt buffer containing 50 mmol/L sodium phosphate (pH 7.0). The cartridge was cooled to 24 °C, and each pad was individually excited by a 635 nm red laser and a 532 nm green laser. The photomultiplier tube used for detection was set at low gain for all three assays, and accumulation times were 1024 µs for factor V, 512 µs for factor II, and 128 µs for MTHFR. Fluorescence emission from the reporter probe dyes was detected at 660–720 nm for Cy5 and 550–600 nm for Cy3.
Data analysis.
Data analysis was performed by the NanoChip Molecular Biology Workstation. The raw fluorescence signals were adjusted by subtracting background signal and normalizing to the heterozygous ratio reference. This signal was reported as adjusted fluorescence for each sample, and the resulting allelic ratios were used to designate the genotype. Mutant/wild-type signal ratios >5:1 were designated homozygous mutants. Wild-type/mutant signal ratios >5:1 were designated homozygous wild type. Ratios <2:1 were designated as heterozygous, and ratios between 5:1 and 2:1 were designated as indeterminate and repeated.
within-cartridge variation
Homozygous mutant, heterozygous, and homozygous wild-type control samples for factor V and factor II were each amplified in 10 separate PCR tubes and loaded to 10 different positions distributed uniformly throughout a single cartridge. Each amplicon of one type was addressed individually to 1 of 10 specified positions during separate portions of the loading program. Fluorescence signals obtained from each of the 10 positions after detection were compared.
between-cartridge variation
Control samples of homozygous mutant, heterozygous, and homozygous wild-type DNA were amplified and loaded with each factor V run. Control data from three runs were compiled and analyzed to evaluate between-cartridge signal variation.
Background signals generated in 10 runs on seven cartridges for factor V and 11 runs on seven cartridges for factor II were also assessed for variation. One additional run was performed for factor II to confirm results
Between-cartridge variation was also evaluated using the fluorescence ratios obtained for heterozygous clinical samples for each of the assays.
minimum template requirements
Homozygous mutant, heterozygous, and homozygous wild-type DNA samples for factor V and factor II were quantified by spectrophotometry (Ultrospec 2100 pro UV/Visible Spectrophotometer; Amersham Biosciences) and diluted to 50, 25, 12, 6, and 3 ng of nucleic acid/µL in 10 mmol/L Tris, 0.1 mmol/L EDTA (pH 8.0). A 2-µL portion of each dilution was PCR amplified and analyzed in the NanoChip platform.
effect of additional mutations
Four clinical DNA samples, two factor II and two MTHFR, were identified through LightCycler testing and ABI Prism 377 (Applied Biosystems) sequencing analysis as containing mutations at sites other than position 20210 for factor II and position 677 for MTHFR. These included one factor II sample that was wild type at 20210 and heterozygous A20218G and a second factor II sample that was wild type at 20210 and heterozygous C20209T. One MTHFR sample was a compound heterozygote C677T/A693T, and a second MTHFR sample was a compound heterozygote C677T/G679A. Each sample was analyzed in the NanoChip platform.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
One factor V sample was reported as indeterminate based on fluorescence signal ratios >2:1 and <5:1. Repeat PCR amplification and detection on a new cartridge resolved the factor V sample. Thirty factor V samples and 19 factor II samples were identified with low signal-to-noise ratios (SNRs) when first tested. One sample repeatedly failed in both the NanoChip and LightCycler systems. One factor V sample and two factor II samples required two repeat amplifications and detections to be resolved. All other samples were resolved after a single repeat PCR amplification and detection.
MTHFR.
DNA from 222 clinical specimens was analyzed in the MTHFR assay on the NanoChip System. No results were obtained for one sample that repeatedly failed in both the NanoChip and LightCycler assays. The agreement between NanoChip results and LightCycler results for MTHFR was 100% (221 of 221). The distribution of samples for MTHFR was 30 homozygous mutant, 90 heterozygous, and 101 homozygous wild type.
Two MTHFR samples were reported as indeterminate based on fluorescence signal ratios >2:1 and <5:1. These samples were resolved by repeat PCR amplification and loading on a new cartridge. Eighteen MTHFR samples were identified with low SNRs when first tested. One sample repeatedly failed in both the NanoChip and LightCycler systems. One of the samples required two repeat amplifications and detections to be resolved, and another sample required three repeats. All other samples were resolved after a single repeat PCR amplification and detection.
signal variation
Within-cartridge signal variation was evaluated using the factor V/factor II multiplex assay for 10 measurements (Fig. 1
). Within-cartridge CVs were <10% when the allele was present and <40% when the allele was not present. The within-cartridge CV for the Cy5 (mutant)/Cy3 (wild type) ratio for the heterozygous samples was 6.9% for factor V and 8.4% for factor II. The fluorescent ratios measured within cartridge for mutant homozygous samples were 26.7–69.5 with a CV of 37% for factor V, and 28.5–43.3 with a CV of 14% for factor II. The fluorescent ratios measured within cartridge for homozygous wild-type samples were 18.3–34.6 with a CV of 21% for factor V and 24.2–53.9 with a CV of 27% for factor II.
|
Between-cartridge variation was examined using the fluorescent signals for the factor V homozygous mutant, heterozygous, and homozygous wild-type control samples from single runs performed on three cartridges. The between-cartridge CV was 11–28%. The range of background signals from 10 runs for factor V on seven cartridges and 11 factor II runs on seven cartridges was 8–34 relative fluorescence units with an overall CV of 38%.
Between-cartridge variation was also examined using the fluorescence ratios for all of the samples with heterozygous results in this study. For 91 factor V heterozygous samples detected in eight runs on seven different cartridges, the mean (SD) fluorescence ratio was 0.87 (0.15) with a CV of 17%. For 36 factor II heterozygous samples detected in nine runs on seven different cartridges, the mean (SD) fluorescence ratio was 0.97 (0.11) with a CV of 11%. For 91 MTHFR heterozygous samples detected in five runs on five different cartridges, the mean (SD) fluorescence ratio was 0.98 (0.17) with a CV of 17%. These data are plotted as the Cy5 (mutant)/Cy3 (wild type) ratio vs the sample number in Fig. 2
.
|
The range of fluorescence ratios for the homozygous wild-type samples in all three assays was 9–7800, and the range for mutant homozygous samples was 7–800.
minimum template requirement
Homozygous mutant, homozygous wild-type, and heterozygous clinical samples for factor V and factor II were tested at 100, 50, 25, 12, and 6 ng of nucleic acid per reaction. The usual input concentration of extracted nucleic acid was 160 ng/reaction. The signals for all dilutions were adequate for determining results, and the appropriate genotypes were assigned in all cases.
effect of additional mutations
Four samples with additional mutations at or near the sequences of factor II and MTHFR probes or stabilizers were evaluated. The locations of the mutations in each of the samples are shown in Fig. 3
. The two factor II samples were correctly identified as wild-type G20210, and one MTHFR sample was correctly identified as heterozygous C677T. The other MTHFR sample was identified as homozygous mutant 677T rather than as heterozygous C677T.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
For the factor V, factor II, and MTHFR assays, a Cy5/Cy3 (mutant/wild-type) fluorescence ratio of 5:1 is required to designate a sample as homozygous mutant, and a Cy3/Cy5 (wild-type/mutant) ratio of 5:1 is required for a designation as homozygous wild type. In this study, the ratios for homozygous samples were generally well above the minimum requirement, and there was little or no difficulty in designating a homozygous sample. The range of ratios for homozygous samples identified in all three assays was 7–7800.
A fluorescence ratio <2:1 was required to designate a sample as a heterozygous genotype. Hybridization of the mutant and wild-type probes to the amplicons from a heterozygous sample should be equivalent. However, even after normalization, variations in the fluorescence emission of the dyes, the lasers, and hybridization contribute to deviation from a precise ratio of 1.0. The data presented in Fig. 2
show that the average heterozygous ratios were 0.87 (factor V), 0.96 (factor II), and 0.98 (MTHFR), indicating that heterozygous samples generated ratios approximating the expected ratio of 1.0. Only one sample from the factor V assay had a fluorescence ratio near 2.0. The Cy5 signal for this sample was comparable to that of other heterozygous samples on the run. The high ratio was a result of a low Cy3 signal. The overall run containing this sample was acceptable, and controls on the run were within specifications. The low Cy3 signal may have been a consequence of the sample location on the microarray, the laser alignment, the sample amplification, or the environment on the microarray pad. The sample was not repeated.
Any sample with a fluorescence ratio between the 5:1 homozygous limit and the 2:1 heterozygous limit was designated as indeterminate and required further analysis. Indeterminate results attributable to fluorescence ratios >2:1 and <5:1 occurred in one factor V sample, zero factor II samples, and two MTHFR samples over the course of this study. All three samples were resolved with repeat PCR amplification and detection on a new cartridge. The incidence of 3 indeterminate results from a total of 681 assays (0.4%) suggests that the genotyping cutoffs are appropriately chosen. Indeterminate frequencies of 0.3% and 0.7% have been noted in other assays (7)(16).
Indeterminate results were also generated when there was a low SNR for a sample. During this study, low SNRs were noted for 
10% of the samples overall. Another report describing the NanoChip System for single-nucleotide polymorphism detection noted a low SNR frequency of 8–12% (17). A total of 67 samples required repeat testing because of low SNRs. All but six samples were resolved on the first repeat testing, which consisted of repeat PCR amplification and detection on a new cartridge. One sample was identified as a failed amplification when it was repeatedly indeterminate in all three assays performed in both the NanoChip and LightCycler systems. Four samples required two repeats to resolve the result, and one sample required repeat testing three times before a result could be obtained. High backgrounds and low signals both affect the SNRs of the assays. Factors contributing to low SNRs can include issues with laser alignment, software versions, suboptimal PCR amplification, amplicon desalting, and background.
Four samples with additional mutations at or near the sequences of the probes or stabilizers used for factor II and MTHFR were evaluated. These samples were identified at ARUP Laboratories based on unusual melting patterns in the factor II and MTHFR LightCycler assays and confirmed by sequence analysis. The locations of these mutations are shown in Fig. 3
.
The first sample was wild-type G20210 for factor II with the unexpected sequence alteration heterozygous A20218G, 8 bases downstream from 20210 and under both the wild-type and mutant probes (Fig. 3A
). This sample was called correctly as wild-type G20210 in the NanoChip system, but the fluorescence signal for the wild-type probe was approximately one-half of the signal for the wild-type control sample. The most likely explanation for the lower signal is that the wild-type probe was binding only to the exact wild-type strand and did not bind to the strand with the A20218G mutation. The mutant probe did not bind to either strand.
The second sample was wild-type G20210 for factor II with the unexpected sequence alteration heterozygous C20209T at the last base at the 5' end of the stabilizer (Fig. 3B
). This sample was correctly called as wild-type G20210 in the NanoChip system. The C20209T mutation at the 5' end of the stabilizer apparently did not affect the stabilizer binding in any way that compromised the detection of the wild-type strand by the wild-type probe.
The third sample was a MTHFR compound heterozygote at C677T and A693G (Fig. 3C
). The A693G mutation was 16 bases down from 677 and 5 bases beyond the probes. This sample was correctly called by the NanoChip system as heterozygous C677T. Because the additional mutation was not under the probes or stabilizer, there was no interference with the NanoChip assay.
The fourth sample was a MTHFR compound heterozygote at C677T and G679A (Fig. 3D
). The G679A mutation was 2 bases down from 677 and under both the wild-type and mutant probes. This sample was called homozygous mutant 677T in the NanoChip system rather than heterozygous C677T. In addition, the mutant probe fluorescence signal was reduced by approximately one-half the expected value. These results suggest that the C677T and G679A mutations are on different strands. The mutant probe was an exact match to the strand that contained only the C677T mutation, but the wild-type probe could not bind to either strand because of the C677T mutation on one strand and the G679A mutation on the other strand.
Any genotyping method that uses probe hybridization risks the possibility of incorrect genotyping from unexpected sequence alterations under the probe. Such incorrect genotyping was noted with a 5'-exonuclease assay for the CETP TaqIB polymorphism (18). In that case, as in the current study, a single hybridization temperature was used for discrimination. The errors in both cases were identified by use of a genotyping assay that performs a melting curve analysis over a range of temperatures. Unexpected sequence alterations that may be missed by hybridization assays detecting at a single temperature can be identified by monitoring hybridization over a range of temperatures. Alternatively, a different gradient of stringency, such as electronic stringency, could be used (15).
Additional mutations near clinically relevant mutation sites are usually rare: for example, in a factor V study only 1 such allele in 4200 was identified (19). However, the ability to assign the appropriate genotype in such samples with additional mutations is important. The NanoChip system is affected by sequence alterations that are under the wild-type or mutant probes, particularly when they are present as compound heterozygotes with the mutation of interest. For example, if the first sample were heterozygous at 20210, it is likely that it would have been incorrectly genotyped. This scenario for factor II may be extremely rare because of the low allele frequency of 20210A in the general population. However, for polymorphisms with higher allelic frequencies, such as for MTHFR, compound heterozygotes may be more frequent. The results presented here also suggest that sequence alterations under the stabilizer may not greatly affect the genotyping results.
In this study, the accuracy of the factor V/factor II and MTHFR assays for the NanoChip Molecular Biology Workstation was supported by the 100% concordance of genotypes obtained with the NanoChip and LightCycler assays on the samples evaluated. The variation of the NanoChip system was acceptable, with within- and between-cartridge CVs for fluorescent signals <10% and 
28%, respectively. The within- and between-cartridge variation of heterozygous ratios measured as CV were 8.4% and 17%, respectively. Accurate results were obtained in this system with input as low as 6 ng of nucleic acid per PCR. An analysis of samples with additional sequence variations near the expected mutations indicated that the presence of such mutations could cause equivocal or incorrect genotype designations. A recent report describing the use of the NanoChip system for the analysis of the factor V mutation also supported the accuracy of the technology (20). Although several samples in that study were discordant when compared with the Third Wave Invader Monoplex Assay, sequence analysis confirmed the NanoChip report.
In conclusion, the factor V, factor II, and MTHFR assays on the NanoChip Molecular Biology Workstation provide accurate and reproducible genotyping. The technical steps of PCR, amplicon desalting, loading, and detection are straightforward. The specific method used for amplicon desalting must be carefully controlled, and the software version for the NanoChip system used in this study was challenging to learn. The multiplex capability and automation of the NanoChip system promise a mechanism for walkaway performance, always useful in a clinical laboratory. A caveat of this system and any direct probe-based method is the possibility of anomalous results when additional sequence variations are present near the mutations of interest, but this is usually a rare occurrence. The overall performance of the NanoChip system indicates promise for routine use in the clinical laboratory.
|
Acknowledgments |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  H.-K. Lee, L. D. Lewis, G. J. Tsongalis, B. C. Schur, P. J. Jannetto, S. H. Wong, and K.-T. J. Yeo Validation of a CYP2D6 Genotyping Panel on the NanoChip Molecular Biology Workstation Clin. Chem., May 1, 2007; 53(5): 823 - 828. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Lyon Discovering Rare Variants by Use of Melting Temperature Shifts Seen in Melting Curve Analysis Clin. Chem., August 1, 2005; 51(8): 1331 - 1332. [Full Text] [PDF]
|
|
|
|
|
|
C. G. Tag, M.-C. Schifflers, M. Mohnen, A. M. Gressner, and R. Weiskirchen Atypical Melting Curve Resulting from Genetic Variation in the 3' Untranslated Region at Position 20218 in the Prothrombin Gene Analyzed with the LightCycler Factor II (Prothrombin) G20210A Assay Clin. Chem., August 1, 2005; 51(8): 1560 - 1561. [Full Text] [PDF]
|
|
|
|
 |
|
 N. A. Saunders, S. Alexander, and I. Tatt env Gene Typing of Human Immunodeficiency Virus Type 1 Strains on Electronic Microarrays J. Clin. Microbiol., April 1, 2005; 43(4): 1910 - 1916. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Bortolin, M. Black, H. Modi, I. Boszko, D. Kobler, D. Fieldhouse, E. Lopes, J.-M. Lacroix, R. Grimwood, P. Wells, et al. Analytical Validation of the Tag-It High-Throughput Microsphere-Based Universal Array Genotyping Platform: Application to the Multiplex Detection of a Panel of Thrombophilia-Associated Single-Nucleotide Polymorphisms Clin. Chem., November 1, 2004; 50(11): 2028 - 2036. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Frusconi, B. Giusti, L. Rossi, S. Bernabini, F. Poggi, I. Giotti, R. Abbate, G. Pepe, and F. Torricelli Improvement of Low-Density Microelectronic Array Technology to Characterize 14 Mutations/Single-Nucleotide Polymorphisms from Several Human Genes on a Large Scale Clin. Chem., April 1, 2004; 50(4): 775 - 777. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
