og D

Pr M
m
E 9J
ra
CM 201

The Journal of Molecular Diagnostics, Vol. 21, No. 6, November 2019

jmd.amjpathol.org

Establishment of Coamplification at Lower

Denaturation Temperature PCR/Fluorescence Melting
Curve Analysis for Quantitative Detection of Hepatitis B
Virus DNA, Genotype, and Reverse Transcriptase Mutation
and Its Application in Diagnosis of Chronic Hepatitis B
Can Liu,*yz Jinpiao Lin,*yz Zhen Xun,*yz Jinlan Huang,*yz Er Huang,*yz Tianbin Chen,*yz
Yujue He,*yz Ni Lin,x Bin Yang,*yz and Qishui Ou*yz

From the Department of Laboratory Medicine,* The First Affiliated Hospital of Fujian Medical University, Fuzhou; the Gene Diagnostic Laboratoryy and the
School of Medical Technology and Engineering,x Fujian Medical University, Fuzhou; and the Fujian Key Laboratory of Laboratory Medicine,z Fuzhou,
People’s Republic of China

CME Accreditation Statement: This activity (“JMD 2019 CME Program in Molecular Diagnostics”) has been planned and implemented in accordance with the accreditation
requirements and policies of the Accreditation Council for Continuing Medical Education (ACCME) through the joint providership of the American Society for Clinical Pathology
(ASCP) and the American Society for Investigative Pathology (ASIP). ASCP is accredited by the ACCME to provide continuing medical education for physicians.

The ASCP designates this journal-based CME activity (“JMD 2019 CME Program in Molecular Diagnostics”) for a maximum of 18.0 AMA PRA Category 1 Credit(s)
. Physicians
should claim only credit commensurate with the extent of their participation in the activity.

CME Disclosures: The authors of this article and the planning committee members and staff have no relevant financial relationships with commercial interests to disclose.

Accepted for publication

August 6, 2019.
Address correspondence to
Qishui Ou, Ph.D., The First
Affiliated Hospital of Fujian
Medical University, 20 Chaz-
hong Rd, Fuzhou, 350005
Fujian, People’s Republic of
China. E-mail: ouqishui@mail.
fjmu.edu.cn.
Dynamic and real-time hepatitis B virus (HBV) DNA, genotype, and reverse transcriptase mutation analysis
plays an important role in diagnosing and monitoring chronic hepatitis B (CHB) and in assessing the
therapeutic response. We established a highly sensitive coamplification at lower denaturation temperature
PCR (COLD-PCR) coupled with probe-based fluorescence melting curve analysis (FMCA) for precision
diagnosis of CHB patients. The imprecision with %CV and detection limit of HBV DNA detected by COLD-
PCR/FMCA were 2.58% to 4.42% and 500 IU/mL, respectively. For mutation, the imprecision and detection
limit were 3.35% to 6.49% and 1%, respectively. Compared with Sanger sequencing, the coincidence rates
of genotype and mutation were 96.0% and 82.5%, respectively, whereas the inconsistent data resulted
from a low proportion (<20%) of mixed genotypes or mixed mutations. The mutation ratio in HBV
infection patients was as follows: hepatitis B e antigen (HBeAg)epositive infection (0/0.0%) < HBeAg-
negative infection (16/4.5%) < HBeAg-positive hepatitis (30/5.5%) < HBeAg-negative hepatitis
(36/6.5%). In patients with entecavir therapy, the proportion of mutation at baseline or week 4 in
virologic response (VR) group was <4%, whereas in the partial VR group, it was mostly 4%. COLD-PCR/
FMCA provides a novel tool with high sensitivity, convenience, and practicability for the simultaneous
quantification of HBV DNA, genotype, and mutation. It might be used for distinguishing the different
phases of HBV infection and predicting VR of CHB patients. (J Mol Diagn 2019, 21: 1106e1116; https://
doi.org/10.1016/j.jmoldx.2019.08.001)

Hepatitis B is still a world-wide prevalent infectious dis-

Supported by National Natural Science Foundation of China grants
ease.1,2 The guidelines confirmed that dynamic monitoring of 81672101 and 81572067 and Joint Funds for the Innovation of Science and
hepatitis B virus (HBV) DNA, genotypes, and reverse tran- Technology of Fujian Province 2016Y9020.
scriptase (RT) mutant DNA is of great importance in the Disclosures: None declared.

Copyright ª 2019 American Society for Investigative Pathology and the Association for Molecular Pathology. Published by Elsevier Inc. All rights reserved.
https://doi.org/10.1016/j.jmoldx.2019.08.001
 COLD-PCR/FMCA for CHB Diagnosis

assessment of infection status, disease progression prediction,
and efficacy judgment of HBV-infected patients.2e4
Currently, several molecular methods have been developed
for detecting HBV genotypes and mutations, including
Sanger sequencing, gene chips, real-time PCR, hybridization-
based assays, next-generation sequencing (NGS), cloning
sequencing, matrix-assisted laser desorption ionization time-
of-flight mass spectrometry, and high-resolution melting.
However, some of these techniques are qualitative detection
with poor sensitivity, some are only for a single mutation, and
the others are expensive and cumbersome to perform with
heavy data analysis workloads and the possibility of nucleo-
tide (nt) mismatches, making them more suitable for research
purposes than clinical application.5,6 Meanwhile, there are
still no high-throughput approaches to detect HBV DNA,
genotype, and RT mutations, simultaneously. Therefore, it is
necessary to establish a more practical and inexpensive
method with high sensitivity to detect genotype and RT mu-
tations while detecting HBV DNA.
Coamplification at lower denaturation temperature PCR
(COLD-PCR) is a novel PCR technique that preferentially
enriches minority alleles from mixtures of wild-type and
mutation-containing sequences.7 The method is based on
changes in the melting temperature (Tm) of double-stranded
DNA containing mismatched base pairs (Figure 1A). The
sensitivity for mutation detection by COLD-PCR can be 10 to
100 times more than that by conventional PCR.8 Probe-based
fluorescence melting curve analysis (FMCA) is a post-PCR
technique that allows detection of multiple mutations by a
single probe based on Tm shifts.9 It is a powerful tool for
mutation detection based on Tm generated by thermal dena-
turation of the probe-target hybrid10 (Figure 1B). COLD-PCR
can be combined with Sanger sequencing strategy to
qualitatively detect HBV genotypic resistance mutations.11
However, the study of quantitative detection of HBV DNA,
genotypes, and RT mutations by COLD-PCR combined with
FMCA has not been reported.
In this study, we established a novel COLD-PCR/FMCA
method that can detect HBV DNA, genotypes, and RT mu-
tations, simultaneously. The analytical performance of this
method, including imprecision, accuracy, sensitivity, detec-
tion limits, linear range, and its ability to detect minor vari-
ants, was systematically evaluated. By analyzing samples
from 650 HBV-infected patients at four different phases and
41 chronic hepatitis B (CHB) patients treated with entecavir,
its clinically diagnostic value to provide precise diagnosis and
treatment for CHB patients was further evaluated.
Materials and Methods
Patients and Study Material
A total of 650 serum specimens were collected from June 5,
2017, to December 1, 2017, at The First Affiliated Hospital of
Fujian Medical University (Fuzhou, China). Patients were all
treatment naive, including 425 outpatients and 225 inpatients.
According to 2017 Clinical Practice Guidelines on the man-
agement of hepatitis B virus infection of European Associa-
tion for the Study of the Liver, HBV infection was further
divided into hepatitis B e antigen (HBeAg)epositive chronic
infection (phase I), HBeAg-positive chronic hepatitis (phase
II), HBeAg-negative chronic infection (phase III), and
HBeAg-negative chronic hepatitis (phase IV).4 Furthermore,
the superinfection or coinfection of hepatitis A, C, D, or E and
HIV had been excluded. The clinical characteristics of the
study population were summarized in Supplemental Table S1.

Figure 1 Schematic workflow for coamplification at lower denaturation temperature PCR (COLD-PCR)/fluorescence melting curve analysis (FMCA).
A: COLD-PCR. Several rounds of conventional PCR cycles produce the initial material for the target amplicons. After denaturation at the critical denaturation
temperature (Tc; eg, 81.5 C), it can preferentially denature strands containing the mutation allele, which generates single-stranded DNA for primer
annealing and extension. B: Probe-based FMCA. After the amplification step, melting curve analysis is conducted by thermal denaturation of the probe-
target hybrid. Different color lines (red and blue lines) represent targets with different mismatches to the hydrolysis probe, with fully matched
wild-type target giving the highest melting temperature value (green line). dF/dT, negative derivative of fluorescent signal corresponding to the
temperature; dsDNA, double-stranded DNA.

The Journal of Molecular Diagnostics - jmd.amjpathol.org 1107

Liu et al

Forty-one CHB patients (21 HBeAg-positive hepatitis
and 20 HBeAg-negative hepatitis patients) receiving ente-
cavir (ETV; 0.5 mg daily) were consecutively recruited and
followed up for 48 weeks from the Liver Research Center of
The First Affiliated Hospital of Fujian Medical University,
between January 2015 and June 2017. Serum samples were
collected at baseline, week 4, week 8, and every 12 weeks
(weeks 12, 24, 36, and 48). On the basis of 2017 Clinical
Practice Guidelines of the European Association for the
Study of the Liver, virologic response (VR) during ETV is
defined as undetectable HBV DNA at 12 months of therapy
in compliant patients. Other cases were defined as partial
virologic responses (PVRs).
The study was approved by the Institutional Medical
Ethics Review Board of The First Affiliated Hospital of
Fujian Medical University (IEC-FOM-013-1.0). Patients all
provided written or verbal consent before inclusion in the
study. Serum samples from the patients were separated and
stored at 80 C for future experiments.
Source of Genomic DNA
For wild-type DNA, pUC-HBV14Kc was used as the tem-
plate (a gift from Lin Xu, Fujian Medical University),
containing full-length HBV DNA genome 3215 bp (Gen-
Bank, https://www.ncbi.nlm.nih.gov/genbank; accession
number AY206380) (genotype B, serotype adw) and
AY206382 (genotype C, serotype adr). For mutant DNAs,
HBV DNA with RT mutations rtA181T, rtA181V,
rtT184I, rtV191I, A194T, rtS202I, rtM204I, rtV207I,
rtL180M þ rtM204V, and rtL180M þ rtM204I (Table 1)
were generated by site-directed mutagenesis using the
corresponding wild-type plasmid as a template; and their
sequences were confirmed by bidirectional DNA
sequencing. The concentration of each recombinant plasmid
DNA was measured by BioPhotometer UV spectropho-
tometer (Eppendorf, Hamburg, Germany).
Conventional PCR Coupled with FMCA
Considering the limitation of the length of the target frag-
ment and the finiteness of the types of mutation recognized
by each probe, if all of the probes were concentrated in a
single tube, it would be difficult to identify the nuances of
Tm values caused by different mutations (Supplemental
Figure S1). Therefore, four probes were designed to
amplify simultaneously in two parallel systems with the
same pair of primers (Table 2). The PCR procedure was
performed in a 25-mL reaction volume containing 50 ng
DNA, 1  Premix Ex Taq (Takara, Dalian, China), 0.4
mmol/L forward primer, 0.04 mmol/L reverse primer in
system A or 0.04 mmol/L forward primer and 0.4 mmol/L
reverse primer in system B (asymmetric PCR), and 0.24
mmol/L probes with lockenucleic acid modification, which
can identify up to 13 known/unknown mutations and
distinguish HBV genotype B or C in each system, in an ABI
7500 real-time fluorescence quantitative PCR instrument
(Applied Biosystems, Foster City, CA).
Conventional PCR cycling conditions were as follows:
95 C for 30 seconds and 45 cycles of 95 C for 10 seconds
and 56 C for 30 seconds. Melting curve analysis began with
a denaturation step of 1 minute at 95 C, a hybridization step
of 1 minute at 40 C, and followed by stepwise temperature
increase from 40 C to 80 C at 0.2 C/second.

Table 1 Partial Sequences of Generating HBV Mutant Plasmids

Genotype Mutation Sequence Position, nt
0 0
B rtA181T 5 -CTCT TGACTCAGTTTACTAGT-3 667w684
rtA181V 50 –CTCT TGGTTCAGTTTACTAGT-30 667w684
rtT184I 50 –CTCT TGGCTCAGTTTATTAGT-30 667w684
rtA194T 50 –GGACTTTCCCCCACTGTC TGG-30 707w727
C rtA181T 50 –CTCC TGACTCAGTTTACTAGT-30 667w684
rtA181V 50 –CTCC TGGTTCAGTTTACTAGT-30 667w684
rtT184I 50 –CTCC TGGCTCAGTTTATTAGT-30 667w684
rtA194T 50 –GGACTTTCCCCCACTGTT TGG-30 707w727
B or C rtL180M 50 –CTCATGGCTCAGTTTACTAGT-30 667w684
rtV191I 50 –AGTGCCATTTGTTCAGTGATT-30 682w702
rtS202I 50 –CATTTATATGGATGATGTGGT-30 732w752
rtM204I 50 –CAGTTATATTGATGATGTGGT-30 732w752
rtM204V 50 –CAGTTATGTGGATGATGTGGT-30 732w752
rtV207I 50 –CAGTTATATGGATGATATTGT-30 732w752
rtL180M þ rtM204V 50 –CTCATGGCTCAGTTTACTAGT-30 667w684
50 –CAGTTATGTGGATGATGTGGT-30 732w752
rtL180M þ rtM204I 50 –CTCATGGCTCAGTTTACTAGT-30 667w684
50 –CAGTTATGTGGATGATGTGGT-30 732w752
The underlined/boldfaced letters represent mutant bases, and the italicized/boldfaced letters represent type-specific bases.
HBV, hepatitis B virus; nt, nucleotide.

1108 jmd.amjpathol.org - The Journal of Molecular Diagnostics

 COLD-PCR/FMCA for CHB Diagnosis

Table 2 The Sequences of Primers and Probes for COLD-PCR/FMCA Detection System
Primers/probes Sequence Position, nt
Primers
Forward 50 -GCCTCAGTCCGTTTCTC-30 650e666
Reverse 50 -AAAGGGACTCAAGATGTT-30 773e790
Probes in system A
FAM-A 50 -FAM-CCAGACAGTGGGGGAAAGCCC-BHQ1-30 707e727
CY5-A 50 -CY5-ACTAGTAAACTGAGCCAG GAG-BHQ2-30 664e684
Probes in system B
FAM-B 50 -FAM-AGTGCCATTTGTTCAGTGGTT-BHQ1-30 682e702
CY5-B 50 -CY5-CAGTTATGTGGATGATGTGGT-BHQ2-30 732e752
The underlined/boldfaced letters represent locked nucleic acidemodified bases, and italicized/boldfaced letters represent type-specific bases.
COLD-PCR, coamplification at lower denaturation temperature PCR; CY5, cyanine 5; FAM, 6-carboxyfluorescein; FMCA, fluorescence melting curve analysis; nt,
nucleotide.

Optimization of COLD-PCR Coupled with FMCA
Fast COLD-PCR was used at a critical denaturation tem-
perature, as previously observed,11 to evaluate the levels of
enrichment of mutant sequences. COLD-PCR was set up
with the same PCR system as described above, at 95 C for
30 seconds, 35 cycles of 95 C for 10 seconds, and 56 C
for 30 seconds (conventional PCR), and 30 cycles of 81.5 C
for 10 seconds and 56 C for 30 seconds (COLD-PCR).
Fluorescence was recorded at 56 C in the corresponding
detection channel. FMCA started with a denaturation step of
1 minute at 95 C, a hybridization step of 1 minute at 40 C,
and followed by stepwise temperature increase from 40 C to
80 C at 0.2 C/second. Fluorescence melting peaks were
built by plotting the negative derivative of fluorescent signal
corresponding to the temperature (dF/dT). Among them,
conventional PCR was used to accurately quantify HBV
DNA, COLD-PCR was used to enrich mutant DNA to
improve detection sensitivity, and FMCA was used to detect
the types and proportion of HBV genotype and mutation.
Sanger Sequencing
The HBV RT gene was amplified by PCR using forward
primer 50 -CTCATGTTGCTGTACAAAACC-30 (nt 559 to
nt 579) and reverse primer 50 -CAATTCKTTGACA-
TACTTTCCA-30 (nt 979 to nt 1000). The PCR conditions
were as follows: 94 C for 5 minutes; 35 cycles of 94 C for 1
minute, 60 C for 1 minute, and 72 C for 1 minute; and then
72 C for 10 minutes. The PCR products were processed for
Sanger dideoxy sequencing with an ABI 3130 genetic
analyzer (Applied Biosystems) with a general PCR purifi-
cation kit (Tiangen Ltd, Beijing, China). To achieve accu-
rate full coverage of each amplicon, PCR products were
sequenced using forward and reverse primers.
Next-Generation Sequencing
NGS was run on a MiSeq sequencer (Illumina, San Diego,
CA) for paired-end sequencing, according to Illumina’s
protocol. To process the raw reads, the quality of raw reads
was evaluated using the online tool fastqc version 0.11.7
(http://www.bioinformatics.babraham.ac.uk/projects/fastqc,
last accessed October 1, 2018). This instrument offers
quantitative single-nucleotide polymorphism and mutation
analyses by rapidly sequencing short stretches of DNA
directly from PCR templates. The following primers were
designed to amplify a 450-bp fragment of the HBV RT
polymerase domain containing positions rt180 to rt215:
forward primer, 50 -GGGCTTTCCCCCACTGTY-30 (nt 708
to nt 725); and reverse primer, 50 -GRGCAACGGGG-
TAAAGGK-30 (nt 1140 to nt 1157). Thirty-five samples,
including 30 clinical samples and 5 plasmid control sam-
ples, were sequenced using the Illumina platform. The
median sequence read length was 406 bp, and the median
coverage was 1500 sequence reads per nucleotide. The
sequence data were analyzed using a quality score threshold
of 30 (corresponding to an error rate of 0.10%), resulting in
shorter and higher standard pair-end reads after quality
control. Therefore, after aligning the sequencing read of the
HBV wild-type plasmid template by NGS with that by
Sanger sequencing, the mean overall error rate was 0.10%
(SD, 0.07%) per base. The frequency threshold was set to 2
SDs higher than the mean error rate (ie, 0.24%), to differ-
entiate a real mutation from a sequencing error.12 RT/S re-
gion sequences were also genotyped with HBV STAR
software version 1.0,13 which is one of the most widely used
software tools for HBV genotyping.
Quantification of HBV DNA
HBV DNA was detected by real-time quantitative PCR
method (Sansure Biotech Inc., Hunan, China) and Roche
Lightcycler 480 (Roche Corp., Basel, Switzerland).
The substance for quality control was from Controls &
Standards Biotechnology (Beijing, China). The lower
limit of detection for HBV DNA was 500 IU/mL. The
linear range was 5.0  102 to 1.0  109 IU/mL of serum
HBV DNA.

The Journal of Molecular Diagnostics - jmd.amjpathol.org 1109

Liu et al

Evaluation and Comparison of COLD-PCR/FMCA with
Other Methods
The evaluation content of the established COLD-PCR/
FMCA detection system included imprecision, accuracy,
limit of detection, linearity, sensitivity, and specificity, ac-
cording to Clinical and Laboratory Standards Institute
EP15-A3, EP09-A3, EP-17A, and EP6-A guidelines.14e17
The magnetic-bead technology was used to extract HBV
DNA from 650 patients in different phases of HBV infection
and 41 CHB patients receiving ETV. Both kits and controls
were provided by Sansure Biotech Inc., as per the manufac-
turer’s instructions. Extracted HBV DNA was subsequently
assayed using COLD-PCR/FMCA, real-time quantitative RT-
PCR from commercial reagents, and conventional PCR/Sanger
sequencing; and NGS methods and results were compared.
Statistical Analysis
Data were analyzed by the SPSS software version 20.0
(IBM, Chicago, IL). Pearson’s c2 test was used to assess the
differences of categorical variables. The t-test and U-test
were used to evaluate the differences of normally distributed
and abnormally distributed continuous variables, respec-
tively. Spearman’s correlation test was used to assess the
relationship between continuous variables. All statistical
tests were two sided, and the level of statistical significance
was set at P < 0.05.
Results
Performance Evaluation of COLD-PCR/FMCA
Imprecision Evaluation
We have successfully established a COLD-PCR/FMCA
method for absolute quantification of HBV DNA and rela-
tive quantification of mutation ratio. According to the Tm
value range of genotype and mutant strains in system A and
system B, HBV genotype B/C and different mutation infection
can be distinguished (Supplemental Table S2). The impreci-
sion of HBV DNA and mutation detection was excellent
(Supplemental Table S3). For quantitative detection of HBV
DNA, the CV fluctuated between 2.58% and 4.42% of within-
laboratory imprecision. For detection of different proportions
of mutant DNA, the CV ranged from 3.35% to 6.49%.
Linearity and Detection Limit Assessment
Serial dilution experiments demonstrated that the amplifi-
cation curves were typically S shaped when the HBV DNA
concentration ranged from 5.0  102 to 5.0  109 IU/mL

Figure 2 The amplification plot and standard curve of coamplification at lower denaturation
temperature PCR/fluorescence melting curve analysis. A: Amplification plot with different colors
represents different concentrations of hepatitis B virus (HBV) DNA. B: Standard curve of absolute
quantification of HBV DNA. C: Melting curve with different colors represents different proportions
of mutant and wild-type DNA. D: Standard curve of relative quantification of mutation ratio.
Eff%, efficiency percentage; Rn, relative fluorescence; CT, cycle threshold.

1110 jmd.amjpathol.org - The Journal of Molecular Diagnostics

 COLD-PCR/FMCA for CHB Diagnosis

(Figure 2A). Standard plasmid DNA was used to establish
the standard curve for detecting HBV DNA (Figure 2B).
The lower detection limit of COLD-PCR/FMCA for
HBV DNA was 500 IU/mL. The mutant DNA was
mixed with the wild-type strains of DNA using serial
dilutions; and the results showed that as the proportion of
mutation increased, the mutant peak gradually increased,
whereas the wild peak gradually decreased and the lower
detection limit for HBV mutation was 1% (Figure 2C). The
observed values of mutation can be calculated by measured
mutation peak/(mutation peak þ wild peak). Standard
curve for relative quantification of mutation ratio was
established using the expected values plotted on the ab-
scissa and the observed values plotted on the ordinate
(Figure 2D).
The Ability to Detect Minor Mutant Subpopulations
A series of mixtures containing plasmids carrying the wild
type and the mutants were prepared. These mixtures were
subjected to conventional PCR/FMCA, COLD-PCR/FMCA,
and PCR Sanger sequencing in parallel. The results showed
that PCR Sanger sequencing could detect the mutant sub-
population when it was present at proportions of 20%. Con-
ventional PCR/FMCA could detect the mutant subpopulation
when it was present at proportions of 10%, whereas COLD-
PCR/FMCA method could detect the subpopulation when it
was present at proportions of 1% (Figure 3, A and B). It
demonstrated that COLD-PCR/FMCA is more sensitive for
the detection of minor mutant subpopulations than conven-
tional PCR/FMCA and PCR Sanger sequencing.
Accuracy Evaluation
The HBV viral concentrations were determined by COLD-
PCR/FMCA and commercial real-time quantitative PCR kit
in these 650 samples. The correlation between the two as-
says was excellent (R2 Z 0.9577, P < 0.05), and the
average deviation was 0.24% (Supplemental Figure S2).
The HBV genotypes and types of mutations were deter-
mined by COLD-PCR/FMCA and Sanger sequencing in the
326 samples in case of patients in which HBV DNA was
1.0  104 IU/mL. The concordance for genotyping be-
tween two methods was 96.0% (313/326; k Z 0.921;
P < 0.001) and for mutation detection was 82.5% (269/326;
k Z 0.223; P < 0.001). The different results were mainly
reflected in a low proportion (<20%) of B/C mixed geno-
types or mixed mutated genes.
Comparison of COLD-PCR/FMCA and NGS
To ensure the accuracy of the COLD-PCR/FMCA results,
especially for minor mutant subpopulations, the relative
quantification of genotype and mutation ratio were detected

Figure 3 A and B: Comparison of conventional PCR/fluorescence melting curve analysis (FMCA), coamplification at lower denaturation temperature PCR
(COLD-PCR)/FMCA, and Sanger sequencing method for minority mutation detection in system. Mutant and wild-type DNAs were diluted in 1:5, 1:10, 1:20,
1:100, and 1:200 serial dilutions. The diluted hepatitis B viruseDNA concentration was 1.0  104 IU/mL. When the rtA181T, rtV191I, rtA194T, or rtM204I
mutation occurred, the sensitivities of conventional PCR/FMCA, COLD-PCR/FMCA, and Sanger sequencing for mutant detection were 10% (1:10), 1% (1:100),
and 20% (1:5), respectively. Black arrows indicate positions of mutated bases.

The Journal of Molecular Diagnostics - jmd.amjpathol.org 1111

Liu et al

Table 3 Comparison of COLD-PCR/FMCA and NGS for Detection of HBV Genotype and Mutation
Mutation(s) detected by COLD-PCR/
No. HBV-DNA, IU/mL Genotype (%) FMCA Nucleotides detected by NGS (%)
1 3.10  105 B (98.15) 2.0% nt724 (C > T) nt724 C (98.15) and T (1.85)
2 4.31  107 C (85.18) 15.0% nt667 (C > T) nt667 C (85.18) and T (14.82)
3 3.33  106 B (98.15) 3.0% nt724 (C > T) nt724 C (97.16) and T (2.84)
4 8.89  104 B (89.84) 9.5% nt724 (C > T) nt724 C (89.84) and T (10.16)
5 3.25  106 C (82.32) 15.0% nt667 (C > T) nt667 C (85.18) and T (14.82)
6 1.34  105 B (98.89) 3.0% nt724 (C > T) þ 3.0% rtA181T nt724 C (98.89) and T (1.11) þ nt670 G
(G > A) (98.64) and A (1.36)
7 6.09  104 C (100) 9.8% rtA181T (G > A) nt670 G (89.24) and A (10.76)
8 3.43  105 B (100) 2.0% rtA181T (G > A) nt670 G (98.78) and A (1.22)
9 8.20  104 B (100) 2.0% rtA181T (G > A) nt670 G (98.15) and A (1.85)
10 1.23  104 B (100) 5.0% rtA181V (C > T) nt671 C (94.62) and T (5.38)
11 3.25  105 C (100) 3.0% rtT184I (C > T) nt680 C (96.72) and T (3.28)
12 1.32  104 C (100) 5.0% rtS202I (G > T) nt734 G (94.64) and T (5.36)
13 1.67  105 C (100) 5.5% rtV207I (G > A) nt748 G (94.06) and A (5.94)
14 5.22  104 C (100) 4.0% rtM204I (G > T) nt741 G (96.23) and T (3.77)
15 5.34  104 C (100) 3.0% rtA181T (G > A) nt667 C (99.34) and A (0.66) þ nt670 G
(99.15) and A (0.85)
16 9.23  106 C (100) 5.5% rtV191I (G > A) þ 5.5% nt700 G (97.67) and A (2.33) þ nt748
rtV207I(G > A) G (97.19) and A (2.91)
17 6.42  104 C (100) 5.0% rtA181T (G > A) þ 5.0% rtV191I nt670 G (98.01) and A (1.99) þ nt700
(G > A) G (97.52) and A (2.48)
18 2.91  104 C (100) 100% rtA180M (G > A) þ 3.5% nt667 C (2.73) and A (97.27) þ nt741 G
rtM204I (G > T) þ 96.5% (98.01) and A (1.99) þ nt739 T
rtM204V (T > G) (1.99) and G (98.01)
19 2.34  104 B (100) 18.5% nt682w702 (unknown) nt688 A (80.5) and G (19.5) (rtI187V)
20 5.97  104 B (100) 3.0% nt707w727 (unknown) nt715 C (97.16) and T (2.84) (rtP196S)
21 4.26  105 B (100) d d
22 6.23  106 B (100) d d
23 4.24  104 B (100) d d
24 3.86  107 B (100) d d
25 6.37  104 C (100) d d
26 1.60  107 C (100) d d
27 2.34  106 C (100) d d
28 2.73  105 C (100) d d
29 2.75  104 B (100) d nt741 G (99.63) and A (0.37)
30 1.14  104 C (100) d nt670 G (99.44) and A (0.56)
The underlined numbers represent the proportion of minor mutant subpopulations.
þ, combination mutations detected; d, no mutation detected; COLD-PCR, coamplification at lower denaturation temperature PCR; FMCA, fluorescence
melting curve analysis; HBV, hepatitis B virus; NGS, next-generation sequencing; nt, nucleotide.

by NGS from 30 samples with wild-type DNA detected by
Sanger sequencing. Among these 30 patients, 22 cases with a
prevalence <20% variation were found by NGS, and COLD-
PCR/FMCA found minor mutant subpopulations in 20 cases.
COLD-PCR/FMCA detected mixed genotypes in six cases
(numbers 1 to 6), single mutation in nine cases (numbers 7 to
approximately 15), combination mutation in three cases
(numbers 16 to 18), and unknown mutation in two cases
(numbers 19 to 20) (Table 3). NGS detected mutations in 22
cases, among which variants present at levels of 1% were
detected by COLD-PCR/FMCA. However, variants present at
levels of <1% were not found by COLD-PCR/FMCA.
The Proportion and Types of Mutation Detected by
COLD-PCR/FMCA Were Higher in Chronic Hepatitis
Patients than Chronic Infection Patients
In the natural history of HBV infection, patients can be
classified into four phases. Four hundred samples of patients
whose HBV DNA was 500 IU/mL at different phases
were analyzed by COLD-PCR/FMCA, and the results
showed that the ratio of genotype C/B in the I through IV
phase group was 0.78:1, 1.30:1, 0.31:1, and 0.52:1,
respectively. The average of HBV-DNA concentration in
genotype C infection group was statistically higher than

1112 jmd.amjpathol.org - The Journal of Molecular Diagnostics

 COLD-PCR/FMCA for CHB Diagnosis

Figure 4 The proportion and types of mutation at different phases of hepatitis B virus infection. A and C: The number and proportion of reverse tran-
scriptase (RT) mutations in genotype B infection patients. B and D: The number and proportion of RT mutations in genotype C infection patients.
***P < 0.001. I, hepatitis B e antigen (HBeAg)epositive chronic infection; II, HBeAg-positive chronic hepatitis; III, HBeAg-negative chronic infection; IV,
HBeAg-negative chronic hepatitis.

genotype B infection group in different phases of HBV phases of HBV infection patients. The number/median
infection patients (Supplemental Table S4). proportion of RT mutations were as follows: phase I
Meanwhile, 318 wild-type and 82 mutation DNAs (0/0.0%) < phase III (16/4.5%) < phase II (30/
were determined by COLD-PCR/FMCA in different 5.5%) < phase IV (36/6.5%) (Figure 4, A and B).

Figure 5 Dynamic changes of the propor-

tion and types of mutation detected by coam-
plification at lower denaturation temperature
PCR/fluorescence melting curve analysis from
baseline to 48 weeks in chronic hepatitis B
patients with virologic response (VR; A and B)
or partial VR (CeE) who received entecavir
therapy. HBV, hepatitis B virus.

The Journal of Molecular Diagnostics - jmd.amjpathol.org 1113

Liu et al
Moreover, 14 combination mutations were detected by
COLD-PCR/FMCA only in the phase of chronic hepa-
titis (phase II or IV) regardless of HBV genotype B or
genotype C infection (Figure 4, C and D). The above
results indicated that COLD/FMCA could distinguish
different phases of HBV infection, according to the
proportion and type of mutations as well as detecting
HBV DNA.
The Genotype and Mutation Detected by COLD/FMCA
Predict Virologic Response of CHB Patients with ETV
Therapy
To evaluate the clinical performance of COLD-FMCA in
the process of antiviral therapy, a total of 41 CHB patients
receiving ETV therapy were enrolled. Of 41 CHB patients,
24 exhibited VR and the other patients showed PVR,
according to HBV DNA concentration detected by COLD-
FMCA (also confirmed by real-time quantitative PCR from
commercial reagents). Fifteen cases of genotype B infection
and nine cases of genotype C infection were detected in the
VR group, whereas eight cases of genotype B infection and
nine cases of genotype C infection were detected in the PVR
group (c2 Z 41.0, P < 0.001). Besides, five cases
(Patients A through E) of RT mutation were detected by
COLD-PCR/FMCA. Patient A had genotype B infection,
with approximately 2.0% of rtA181T mutation at baseline,
at week 4 and week 8 (Figure 5A); and Patient B had ge-
notype C infection, with approximately 2.0% of rtA181T
mutation at baseline and week 4 (Figure 5B). The other
three cases of CHB patients with PVR were genotype
C infection. Approximately 6.0% of rtA181T mutation was
detected in Patient C from baseline to week 48 (Figure 5C).
Of rtA181T mutation, 2% was detected at baseline and
4% was detected from week 4 to week 48 in patient D
(Figure 5D). Of rtM204I mutation, 6% to 20% was detected
in Patient E from baseline to week 48 (Figure 5E). All these
results suggested that the dynamic changes of HBV DNA,
genotype, and mutation detected by COLD-PCR/FMCA
may contribute to more precisely diagnose and predict VR
of CHB patients during ETV therapy.
Discussion
Dynamic and real-time HBV DNA, genotype, and RT
mutation analysis plays an important role in diagnosing and
monitoring HBV infection as well as in assessing the ther-
apeutic response. At present, there are various limitations of
other methods that can be used to detect these indicators,
and there is still no high-throughput system for their
simultaneous detection. COLD-PCR technology has been
widely used in the detection of mutations because of its high
sensitivity, simplicity, economy, and practicality.18e21
FMCA technology is also widely used in the detection of
macrolide resistance, thalassemia genetic screening, and
1114
JAK2/MPL mutation.22e24 In the previous experiments,11
COLD-PCR was first combined with Sanger sequencing
for qualitative detection of HBV genotypic resistance mu-
tations. However, the application of COLD-PCR with
FMCA method for the diagnosis and treatment of HBV
infection has not been reported.
In this study, first, COLD-PCR/FMCA technology was
successfully established to achieve accurate quantification
of HBV DNA concentrations in the range of 5.0  102 to
5.0  109 IU/mL, with a detection limit of 5.0  102
IU/mL. At present, most HBV genotyping and mutation
detection platforms, such as Sanger sequencing and NGS
technology, require at least HBV DNA 1.0  104 IU/mL
for analysis. However, studies have shown that the mutation
rate in the S/RT region is significantly increased when HBV
DNA is <2.0  104 IU/mL.25 These results also showed
that 20.7% (17/82) of the mutant DNA concentration was
between 5.0  102 and 1.0  104 IU/mL. Therefore, a good
limit of detection allowed mutations in HBV-infected pa-
tients to be recognized at lower viral loads. Second, this
system can accurately determine the B/C genotype by
FMCA. Although HBV can be classified into 10 genotypes
of A to J, genotypes B and C were the most predominant
genotypes in China.26,27 Consistently, a large number of
clinical sample tests in our laboratory had also shown that
the current prevalence of HBV genotype is B or C.28 The
HBV genotype results, detected by type-specific probes in
this study, were 96.0% consistent with Sanger sequencing
results. There was a difference between the two methods in
the detection of low-proportion B/C mixed genotypes, as
confirmed by NGS. Furthermore, this method takes advan-
tage of the high sensitivity of COLD-PCR for mutation
detection, which greatly improves the detection ability of
minor mutant subpopulations (up to 1%). This result was 20
times more sensitive than Sanger sequencing, which is
similar to the findings from Pang et al29 for detection of
IDH1 mutations. Compared with the gradient approach
(full COLD-PCR) published by Galbiati et al,30 the fast
COLD-PCR can quantitatively measure the mutation, which
is more suitable to monitor the dynamic process of HBV
infection. This study showed that approximately 17.5% of
the mixed mutations were detected by the COLD-PCR/
FMCA method, whereas no mutations were detected by
Sanger sequencing, which was confirmed to be low-
prevalence mutations by NGS analysis. It is interesting
that some mutations were only detected by NGS but not by
COLD-PCR. This may be because of the low levels of
mutations (<1.0%) in the samples, which were below the
detection limit of COLD-PCR. In the follow-up study,
COLD-PCR will be combined with other more sensitive
technology, such as digital PCR,31,32 to improve the
detection limit of mutation and further confirm the differ-
ences between NGS and COLD-PCR. Therefore, the novel
method is accurate, stable, sensitive, and suitable for
quantitative monitoring of HBV DNA, genotypes, and RT
mutations in HBV infection patients.
jmd.amjpathol.org - The Journal of Molecular Diagnostics

According to 2017 Clinical Practice Guidelines of the
European Association for the Study of the Liver, the nat-
ural history of HBV infection can be classified into four
phases. However, even after a complete assessment, some
subjects fall into an indeterminate gray area, which results
in the dilemma in antiviral therapy.33 COLD-PCR/FMCA
technology was used to detect clinical specimens of
different HBV infection phases, the proportion of genotype
C in the HBeAg-positive group was found to be signifi-
cantly higher than that in HBeAg-negative group, and
genotype C was more likely to progress to hepatitis than
genotype B. Meanwhile, detection of RT mutation ratios
showed the following: phase I < phase III < phase
II < phase IV. And the proportion of mutations in 90%
HBV infection patients was <20%, which fully confirms
that it is necessary to use highly sensitive techniques for
mutation detection. It could be distinguished between
phase III and IV when the 5.0% mutation ratio was used as
the diagnostic cutoff point. This research also found that
combination mutations can only occur in the hepatitis
patients (phase II or IV) who required antiviral therapy.
This may be because of the fact that at phase I or phase
III, the host exerts less immune pressure and the virus
replicates in a more homogeneous form. Once the pro-
gression is to phase II or phase IV, HBV replication is
greatly affected by the external environment and host
immunity, and the heterogeneity of HBV RT quasispecies
begins to increase.34,35 Therefore, it is suggested that the
disease was more inclined to a severe hepatitis phase when
a mutation was detected by the newly established method,
especially a higher ratio (5.0%) or combination mutation.
Moreover, in CHB patients with ETV therapy for 48 or
52 weeks, the virologic response rates are 60% to 90%.4,36
Current guidelines suggest that quantitative HBV DNA and
hepatitis B surface antigen have a predictive value in the
antiviral therapeutic efficacy that happens 24 or 48 weeks
later.3,4 This study also used COLD-PCR/FMCA to detect
serum samples during antiviral therapy. The results showed
that the genotype B was more susceptible to VR than ge-
notype C. CHB patients who received ETV therapy with
4% of RT mutation at baseline or week 4, detected by
COLD-PCR/FMCA, may be more prone to develop PVR,
which suggested that with high barrier to resistance nucle-
os(t)ide analogues, such as ETV, there were few clinical
cases that directly lead to genotypic resistance. However,
the quasispecies characteristics of HBV replication deter-
mine that its heterogeneity will still vary. Under the pressure
of host and drug, resistant strains with single-point
mutations are difficult to progress to dominant pop-
ulations, but evolved in the form of low-abundance or
multisite combination mutations.37e39 However, limited by
the number of samples tested, the prediction value can be
only set to 4.0% initially. With the addition of subsequent
sample sizes and further research, the clinical value of
COLD-PCR/FMCA technology in antiviral therapy will be
further improved.
The Journal of Molecular Diagnostics - jmd.amjpathol.org
COLD-PCR/FMCA for CHB Diagnosis
In summary, we successfully established a highly sensi-
tive COLD-PCR/FMCA method for the simultaneous
quantification of HBV DNA, genotype, and mutation. It is
simple, convenient, practical, and inexpensive and can be
performed routinely in most grass-roots hospital labora-
tories. It provides a powerful laboratory tool for precise
diagnosis and treatment of HBV infection patients.
Acknowledgments
We thank colleagues, especially Dr. Jing Chen, at The First
Affiliated Hospital of Fujian Medical University for support
of this study. pUC-HBV14Kc was a gift from Lin Xu
(Fujian Medical University).
Supplemental Data
Supplemental material for this article can be found at
http://doi.org/10.1016/j.jmoldx.2019.08.001.
References
1. Ott JJ, Stevens GA, Groeger J, Wiersma ST: Global epidemiology of
hepatitis B virus infection: new estimates of age-specific HBsAg
seroprevalence and endemicity. Vaccine 2012, 30:2212e2219
2. Chinese Society of Hepatology CMA, Chinese Society of Infectious
Diseases: The guideline of prevention and treatment for chronic
hepatitis B: a 2015 update. Zhonghua Gan Zang Bing Za Zhi 2015,
23:888e905
3. WHO Guidelines Approved by the Guidelines Review Committee:
Guidelines for the Prevention, Care and Treatment of Persons with
Chronic Hepatitis B Infection. Geneva, Switzerland, World Health
Organization, 2015
4. European Association for the Study of the Liver: EASL 2017 Clinical
Practice Guidelines on the management of hepatitis B virus infection.
J Hepatol 2017, 67:370e398
5. Rybicka M, Stalke P, Bielawski KP: Current molecular methods for
the detection of hepatitis B virus quasispecies. Rev Med Virol 2016,
26:369e381
6. Kim JH, Park YK, Park ES, Kim KH: Molecular diagnosis and
treatment of drug-resistant hepatitis B virus. World J Gastroenterol
2014, 20:5708e5720
7. Li J, Wang L, Mamon H, Kulke MH, Berbeco R, Makrigiorgos GM:
Replacing PCR with COLD-PCR enriches variant DNA sequences
and redefines the sensitivity of genetic testing. Nat Med 2008, 14:
579e584
8. Milbury CA, Correll M, Quackenbush J, Rubio R, Makrigiorgos GM:
COLD-PCR enrichment of rare cancer mutations prior to targeted
amplicon resequencing. Clin Chem 2012, 58:580e589
9. Elenitoba-Johnson KS, Bohling SD, Wittwer CT, King TC: Multiplex
PCR by multicolor fluorimetry and fluorescence melting curve anal-
ysis. Nat Med 2001, 7:249e253
10. Huang Q, Liu Z, Liao Y, Chen X, Zhang Y, Li Q: Multiplex fluo-
rescence melting curve analysis for mutation detection with dual-
labeled, self-quenched probes. PLoS One 2011, 6:e19206
11. Liu C, Lin J, Chen H, Shang H, Jiang L, Chen J, Ye Y, Yang B,
Ou Q: Detection of hepatitis B virus genotypic resistance mutations
by coamplification at lower denaturation temperature-PCR coupled
with Sanger sequencing. J Clin Microbiol 2014, 52:2933e2939
12. Yang G, Liu Z, Yang J, Luo K, Xu Y, He H, Fu Q, Yu S, Wang Z:
Quasispecies characteristics in mother-to-child transmission of
1115
Liu et al
hepatitis B virus by next-generation sequencing. J Infect 2017, 75:
48e58
13. Myers R, Clark C, Khan A, Kellam P, Tedder R: Genotyping hepa-
titis B virus from whole- and sub-genomic fragments using position-
specific scoring matrices in HBV STAR. J Gen Virol 2006, 87:
1459e1464
14. CLSI: User Verification of Precision and Estimation of Bias:
Approved Guideline-Third Edition CLSI Document EP15-A3.
Wayne, PA, Clinical and Laboratory Standards Institute, 2014
15. CLSI: Measurement Procedure Comparison and Bias Estimation
Using Patient Sample: Approved Guideline-Third Edition. Wayne,
PA, Clinical and Laboratory Standards Institute, 2013
16. CLSI: Evaluation of the Linearity of Quantitative Measurement
Procedures: A Statistical Approach: Approved Guideline CLSI
Document EP06-A. Wayne, PA, Clinical and Laboratory Standards
Institute, 2003
17. CLSI: Evaluation of Detection Capability for Clinical Laboratory
Measurement Procedures: Approved Guideline-Second Edition CLSI
Document EP17-A2. Wayne, PA, Clinical and Laboratory Standards
Institute, 2012
18. Castellanos-Rizaldos E, Richardson K, Lin R, Wu G,
Makrigiorgos MG: Single-tube, highly parallel mutation enrichment
in cancer gene panels by use of temperature-tolerant COLD-PCR.
Clin Chem 2015, 61:267e277
19. Mauger F, Daunay A, Deleuze JF, Tost J, How-Kit A: Multiplexing
of E-ice-COLD-PCR assays for mutation detection and identification.
Clin Chem 2016, 62:1155e1158
20. Galbiati S, Monguzzi A, Damin F, Soriani N, Passiu M, Castellani C,
Natacci F, Curcio C, Seia M, Lalatta F, Chiari M, Ferrari M,
Cremonesi L: COLD-PCR and microarray: two independent highly
sensitive approaches allowing the identification of fetal paternally
inherited mutations in maternal plasma. J Med Genet 2016, 53:
481e487
21. Mauger F, How-Kit A, Tost J: COLD-PCR technologies in the area of
personalized medicine: methodology and applications. Mol Diagn
Ther 2017, 21:269e283
22. Touati A, Peuchant O, Jensen JS, Bebear C, Pereyre S: Direct
detection of macrolide resistance in Mycoplasma genitalium isolates
from clinical specimens from France by use of real-time PCR and
melting curve analysis. J Clin Microbiol 2014, 52:1549e1555
23. Ahn JJ, Kim Y, Lee SY, Hong JY, Kim GW, Hwang SY: Fluores-
cence melting curve analysis using self-quenching dual-labeled pep-
tide nucleic acid probes for simultaneously identifying multiple DNA
sequences. Anal Biochem 2015, 484:143e147
24. Park J, Song M, Jang W, Chae H, Lee GD, Kim K, Park H, Kim M,
Kim Y: Peptide nucleic acid probe-based fluorescence melting curve
analysis for rapid screening of common JAK2, MPL, and CALR
mutations. Clin Chim Acta 2017, 465:82e90
25. Mirabelli C, Surdo M, Van Hemert F, Lian Z, Salpini R, Cento V,
Cortese MF, Aragri M, Pollicita M, Alteri C, Bertoli A, Berkhout B,
Micheli V, Gubertini G, Santoro MM, Romano S, Visca M,
Bernassola M, Longo R, De Sanctis GM, Trimoulet P, Fleury H,
1116
Marino N, Mazzotta F, Cappiello G, Spano A, Sarrecchia C, Zhang JM,
Andreoni M, Angelico M, Verheyen J, Perno CF, Svicher V: Specific
mutations in the C-terminus domain of HBV surface antigen signifi-
cantly correlate with low level of serum HBV-DNA in patients with
chronic HBV infection. J Infect 2015, 70:288e298
26. Tian Q, Jia J: Hepatitis B virus genotypes: epidemiological and
clinical relevance in Asia. Hepatol Int 2016, 10:854e860
27. Lin CL, Kao JH: Hepatitis B virus genotypes and variants. Cold
Spring Harb Perspect Med 2015, 5:a021436
28. Wang W, Liang H, Zeng Y, Lin J, Liu C, Jiang L, Yang B, Ou Q:
Establishment of a novel two-probe real-time PCR for simultaneously
quantification of hepatitis B virus DNA and distinguishing genotype
B from non-B genotypes. Clin Chim Acta 2014, 437:168e174
29. Pang B, Durso MB, Hamilton RL, Nikiforova MN: A novel COLD-
PCR/FMCA assay enhances the detection of low-abundance IDH1
mutations in gliomas. Diagn Mol Pathol 2013, 22:28e34
30. Galbiati S, Brisci A, Lalatta F, Seia M, Makrigiorgos GM, Ferrari M,
Cremonesi L: Full COLD-PCR protocol for noninvasive prenatal
diagnosis of genetic diseases. Clin Chem 2011, 57:136e138
31. Gullett JC, Nolte FS: Quantitative nucleic acid amplification methods
for viral infections. Clin Chem 2015, 61:72e78
32. Liu Y, Cathcart AL, Delaney WEt, Kitrinos KM: Development of
a digital droplet PCR assay to measure HBV DNA in patients
receiving long-term TDF treatment. J Virol Methods 2017, 249:
189e193
33. Yang HC, Kao JH: Looking into the crystal ball: biomarkers for
outcomes of HBV infection. Hepatol Int 2016, 10:99e101
34. Li H, Song XF, Hu TT, Ren H, Hu P: A strong conservative tendency
in HBV transcriptase (RT): a majority of natural RT mutations
derived from the S gene. Liver Int 2016, 36:963e970
35. Yang ZT, Huang SY, Chen L, Liu F, Cai XH, Guo YF, Wang MJ,
Han Y, Yu DM, Jiang JH, Zhang DH, Gong QM, Zhang GQ,
Zang GQ, Lu ZH, Huang LH, Zhang XX: Characterization of full-
length genomes of hepatitis B virus quasispecies in sera of patients
at different phases of infection. J Clin Microbiol 2015, 53:
2203e2214
36. Pan HY, Pan HY, Song WY, Zheng W, Tong YX, Yang DH,
Dai YN, Chen MJ, Wang MS, Huang YC, Zhang JJ, Huang HJ:
Long-term outcome of telbivudine versus entecavir in treating higher
viral load chronic hepatitis B patients without cirrhosis. J Viral
Hepatitis 2017, 24 Suppl 1:29e35
37. Liu F, Chen L, Yu DM, Deng L, Chen R, Jiang Y, Chen L,
Huang SY, Yu JL, Gong QM, Zhang XX: Evolutionary patterns of
hepatitis B virus quasispecies under different selective pressures:
correlation with antiviral efficacy. Gut 2011, 60:1269e1277
38. Gao S, Duan ZP, Coffin CS: Clinical relevance of hepatitis B virus
variants. World J Hepatol 2015, 7:1086e1096
39. Zhang HY, Liu LG, Ye CY, Chen CH, Hang SX, Zhu Z,
Shen HY, Huang ZY, Chen WY, Xue Y: Evolution of drug-
resistant mutations in HBV genomes in patients with treatment
failure during the past seven years (2010-2016). Virus Genes 2018,
54:41e47
jmd.amjpathol.org - The Journal of Molecular Diagnostics