Comparison of Third-Generation Sequencing and Routine Polymerase Chain Reaction in Genetic Analysis of Thalassemia

Subjects from Hunan Province were recruited at Changsha Hospital for Maternal & Child Health Care Affiliated to Hunan Normal University (Changsha, China). Hb testing was conducted for these subjects. CATSA and routine PCR were simultaneously performed for 504 Hb testing–positive subjects. Ethics approval was given by the institutional review board (No. 2021097). All subjects provided informed written consent.

Hb testing was carried out by standard blood assays and Hb electrophoresis. Standard blood assays were conducted with an automated cell counter (Sysmex XN-1000, Sysmex Co, Ltd, Kobe, Japan). Blood indexes included mean corpuscular volume, mean corpuscular Hb, and Hb. An Hb electrophoresis system (Capillarys 2 Flex Piercing, Sebia, Evry Cedex, France) was used to analyze the Hb components, which included HbA₂ and HbF. Normal ranges were mean corpuscular volume 80 fL or higher, mean corpuscular Hb 27 pg or higher, HbA₂ levels between 2.5% and 3.5%, and HbF 5% or lower.

PCR was performed to test α-thalassemia variants including –^SEA (Southeast Asia), -α^3.7 (rightward), -α^4.2 (leftward), HBA2:c.427T>C, HBA2:c.377T>C, and HBA2:c.369C>G and β-thalassemia variants including HBB:c.126_129delCTTT, HBB:c.130G>T, HBB:c.316-197C>T, HBB:c.52A>T, HBB:c.45_46insG, HBB:c.-76A>G, HBB:c.-79A>G, HBB:c.216_217insA, HBB:c.79G>A, HBB:c.92+1G>T, HBB:c.92+1G>A, HBB:c.84_85insC, HBB:c.92+5G>C, HBB:c.-11_-8delAAAC, HBB:c.-50A>C, HBB:c.2T>G, HBB:c.94delC, HBB:c.-80T>C, and HBB:c.-82C>A. The experiment was performed according to the manufacturer’s protocol (Kaipu Bioscience, Chaozhou, China).

CATSA was performed at Berry Genomics as described previously.²⁴  Briefly, multiplex PCR was conducted to amplify the genomic regions covering the majority of known structural variations, single-nucleotide variants (SNVs), and indels for the HBA1, HBA2, and HBB genes. Bar-coded adaptors were ligated to the PCR products by a 1-step end-repair and ligation reaction. Unligated products were removed by exonucleases (Enzymatics), and libraries were pooled together by equal mass. The pooled library was converted to SMRT bell library by Sequel Binding and Internal Ctrl Kit 3.0 (Pacific Biosciences) and sequenced with Sequel II Sequencing Kit 2.0 using the CCS mode on the Sequel II platform (Pacific Biosciences). The raw reads were converted to CCS reads by CCS software (Pacific Biosciences) and demultiplexed by bar codes using lima in the Pbbioconda package (Pacific Biosciences). Then, the processed reads were then aligned to genome build hg38 with pbmn2. FreeBayes1.3.4 (https://www.geneious.com/plugins/freebayes; Biomatters Inc, San Diego, California) was used to call SNVs, and indels and structural variations were identified based on HbVar, Ithanet, and LOVD databases.

Discordant structural variants identified by CATSA but missed by routine PCR were verified by specially designed PCR. Discordant SNVs and indels between CATSA and routine PCR were validated by Sanger sequencing.

Subjects in Hunan Province were recruited and hematologic testing was performed. CATSA and routine PCR were conducted for 504 Hb testing–positive subjects side by side (Figure 1; see Supplemental Table in the supplemental digital content, containing 1 table and 3 figures, at https://meridian.allenpress.com/aplm in the March 2024 table of contents). Of these subjects, 462 (91.67%) had the same results whereas 42 (8.33%) exhibited discordant results between the 2 methods. Among the 42 discordant results, 4 variants were within the PCR detection range (common thalassemia variants), and 38 variants were outside the PCR detection range. Of those variants outside the PCR detection range, 19 were rare thalassemia variants and 19 had unknown pathogenicity. Sanger sequencing and specially designed PCR confirmed that the results of CATSA were all correct. In total, CATSA correctly detected 247 subjects with variants, whereas PCR correctly identified 205 subjects with variants, which showed an increase of 20.49%.

Among 228 subjects diagnosed as carriers of thalassemia with common or rare variants by CATSA, 103 (45.18%) were diagnosed as carriers of α-thalassemia, 110 (48.25%) were carriers of β-thalassemia, and 15 (6.58%) had coinheritance of α- and β-thalassemia (Table 1). Among the 103 α-thalassemia carriers, 62 (60.19%) carried –^SEA/αα, which was the most common genotype. The next 4 most common genotypes were -α^3.7/αα (12.62%), -α^4.2/αα (5.83%), HBA2:c.427T>C Hete (3.88%), and HBA2:c.369C>G Hete (2.91%). Among the 110 β-thalassemia carriers, HBB:c.316-197C>T Hete was the most common genotype (39 subjects; 35.45%). The next 4 most common genotypes were codons HBB:c.126_129delCTTT Hete (25.45%), HBB:c.52A>T Hete (15.45%), HBB:c.216_217insA Hete (4.55%), and HBB:c.84_85insC Hete (3.64%). In the 15 α+β-thalassemia carriers, there were 14 different genotypes. Two subjects had the genotype of -α^3.7/αα compound with HBB:c.316-197C>T Hete.

Of the 504 subjects, 4 subjects had discordant results between routine PCR and CATSA within the PCR detection range (Table 2). Sanger sequencing and specially designed PCR verified the results of the CATSA were correct (Supplemental Figure 1). Sequencing reads generated by CATSA for discordant variants were displayed in Integrative Genome Viewer (Figure 2, A through D). Two subjects had discordant variants in HBA genes, and 2 had discordant variants in HBB genes. For subject 21KDP33313, PCR could not determine the correct variant but CATSA confirmed she did not possess -α^3.7. For subjects 21KDP33003, 21KDP33032, and 21KDP33371, the routine screening protocol tested only HBA genes or HBB genes by PCR based on the results of the hematologic testing. The results of PCR were normal, whereas CATSA identified thalassemia variants that were within the PCR detection range. Subject 21KDP33371 had hypochromic microcytosis, and subject 21KDP33003 had moderate anemia. The results of CATSA were more consistent with their phenotypes. To further confirm the results of CATSA, routine PCR was performed again for the missing alleles. The results of the second round of PCR were the same with CATSA.

In addition to the thalassemia variants tested by routine PCR, CATSA can also detect rare thalassemia variants in α- and β-globin genes. Of the 504 subjects, 19 subjects had rare thalassemia variants detected by CATSA (Table 3). Sanger sequencing and specially designed PCR verified that the results of the CATSA were correct (Supplemental Figure 2). Sequencing reads generated by CATSA of the representative variants were displayed in Integrative Genome Viewer (Figure 2). Eighteen subjects had discordant variants in HBA genes, and 1 had a discordant variant in HBB genes. Among these 18 subjects, 10 subjects had additional α triplications identified by CATSA and 8 subjects had additional insertions, deletions, and SNVs detected by CATSA. In the 10 subjects with α triplications, 6 were also compounded with β-thalassemia. Most of the 6 subjects with α triplications and β-thalassemia variants had hypochromic microcytosis and anemia. For subject 21KDP33130, the genotype determined by PCR was HBB:c.316-197C>T Hete, whereas CATSA identified additional ααα^anti3.7/αα (Figure 2, E). Her Hb was only 92 g/L when she was 6 months old, which was more consistent with the result of CATSA. For subject 21KDP33235, PCR detected only HBB:c.126_129delCTTT Hete, whereas CATSA identified additional ααα^anti4.2/αα (Figure 2, F). She suffered from hypochromic microcytosis and her Hb level was only 98 g/L. For subject 21KDP33087, the genotype determined by PCR was αα/-α^4.2, which predicted silent thalassemia. However, that determined by CATSA was -α^4.2/HBA1:c.223G>C, which predicted thalassemia trait. For subject 21KDP33442, the genotype determined by PCR was αα/-α^3.7 whereas that determined by CATSA was -α^3.7/HBA1:c.364G>A. This subject had hypochromic microcytosis, low HbA₂, and mild anemia, which were more consistent with the result of CATSA. For subjects 21KDP33033, 21KDP33075 (Figure 2, G), 21KDP33206, 21KDP33339, and 21KDP33472 (Figure 2, I), the results of PCR were normal, whereas CATSA detected rare thalassemia SNVs and indels. These subjects had either hypochromic microcytosis or low HbA₂, which was more consistent with the result of CATSA. For subject 21KDP33057, PCR only detected 1 thalassemia variant in the HBB gene, whereas CATSA detected 2 thalassemia variants in the HBB gene. Particularly, CATSA revealed that the variants were in trans configurations (Figure 2, H), which predicted β-thalassemia intermedia/major.

Nineteen subjects had discordant results of unknown pathogenicity between CATSA and PCR (Table 4; Figure 2). Sanger sequencing confirmed that the results of CATSA were correct (Supplemental Figure 3). Five subjects had discordant SNVs in HBA genes. Fourteen subjects had discordant SNVs in HBB genes. Among these discordant variants, the most frequent one was HBB:c.315+180T>C Hete.

In the subjects with positive results for Hb testing, 9 did not possess thalassemia variants but had other SNVs in their HBA and HBB genes. Three subjects had SNVs in HBA genes. Six subjects had SNVs in HBB genes, among whom 3 harbored HBB:c.315+180T>C Hete. Because these subjects all exhibited at least 1 abnormal hematologic parameter, these 7 SNVs may have potential pathogenicity of thalassemia (Table 5).

In this study, the comparison of CATSA and routine PCR of samples with positive results for Hb testing was carried out on a large scale. Among the 504 subjects, 462 (91.67%) had the same results between CATSA and PCR, whereas 42 (8.33%) had discordant results. Specially designed PCR or Sanger sequencing verified the results of CATSA. In this study, routine PCR had several false-positive and false-negative results. PCR could not detect -α^3.7 and did not detect α triplications in 10 subjects (of 504; 1.98%). The incidence of ααα^anti3.7 and ααα^anti4.2 was as high as 1% to 2% among the general population in southern China.¹³  Because α-globin triplications can cause β-thalassemia intermedia when compounded with β⁰ or β⁺ heterozygotes, it would be critical to include these variants in thalassemia carrier screening and prenatal diagnosis. In addition, routine PCR missed rare SNVs and indels, and other Hb variants in 9 additional subjects. Because these subjects all exhibited at least 1 abnormal hematologic parameter, these SNVs may have potential pathogenicity of thalassemia.

Recently, the clinical utility of TGS in the diagnosis of thalassemia was validated, and TGS was demonstrated as a novel and powerful tool in the genetic analysis of thalassemia. Zhuang et al²⁷  performed TGS in 70 suspected carriers of rare thalassemia variants and found that TGS yielded a 7.14% (5 of 70) increment of rare α- and β-globin gene variants as compared with PCR and Sanger sequencing. Peng et al²⁸  conducted TGS for 100 cases of suspected thalassemia and identified an additional 10 cases of rare thalassemia variants compared with traditional thalassemia testing. Long and Liu¹⁴  carried out TGS in 4 samples that showed abnormalities in the traditional genetic tests and demonstrated that they carried complex α-globin gene variants. Jiang et al²⁹  conducted TGS in 9 samples that showed abnormalities in the traditional genetic tests and found that they possessed various complex variants. However, the comparison between TGS and routine methods in Hunan Province has not been explored.

In China, the routine strategy for thalassemia screening is Hb testing. Then, PCR for one or both of the HBA and HBB genes is performed based on the result of Hb testing. In this research, 8 subjects did not have their HBA genes and 280 subjects did not have their HBB genes tested by routine PCR. However, CATSA found that 1 of the 8 subjects possessed an HBA variant within the PCR detection range and 2 of the 280 subjects possessed HBB variants within the PCR detection range. To further confirm the results of CATSA, routine PCR was performed again for the missing alleles in 288 subjects. The results of the second round of routine PCR were the same with CATSA. This means all the HBA and HBB alleles should be tested by genetic analysis.

In conclusion, CATSA is a more comprehensive, reliable, and effective approach in genetic analysis of thalassemia compared with PCR. In terms of cost, previous research^30,31  demonstrated that TGS based on the PacBio platform could competitively deliver a test as low as around $20. However, there are still potential drawbacks of CATSA relative to other methods. The PacBio sequencing platform Sequel II is expensive and hundreds of samples need to be pooled together for sequencing in one flow cell to reduce the cost per sample, which makes it suitable only for large centers at present. Thus, developing a benchtop PacBio sequencing platform with lower cost and lower throughput would be more clinically feasible. Currently, large deletions are often identified by multiple-ligation probe amplification, which determines the copy number of each probe using quantitative analysis. Unlike multiple-ligation probe amplification, CATSA relies on the combination of primers to directly identify the breakpoints of deletions, which is more straightforward but might miss novel deletions that are not covered by the primer set. Thus, it is important to review and optimize the primer set of CATSA periodically for comprehensive coverage of rare deletions. There are also certain limitations in the present study. First, the current method only covered HBA and HBB genes. In the future, it is technically possible to include other clinically significant regions, such as the HS40 region and modifier genes. Second, this study evaluated the clinical utility of CATSA only in samples with positive results for Hb testing. Future experiments can be conducted in Hb testing–negative samples to compare the effectiveness of PCR and CATSA. Third, this was a retrospective study with a limited sample size. A large-scale and multicenter prospective study can be directed to fully explore the molecular spectrum of the subjects in Hunan Province.