The third generation sequencing: the advanced approach to genetic diseases.

Genomic sequencing technologies have revolutionized mutation detection of the genetic diseases in the past few years. In recent years, the third generation sequencing (TGS) has been gaining insight into more genetic diseases owing to the single molecular and real time sequencing technology. This paper reviews the genomic sequencing revolutionary history first and then focuses on the genetic diseases discovered through the TGS and the clinical effects of the TGS, which is followed by the discussion of the improvement in the bioinformatic analysis for the TGS and its limitations. In summary, the TGS has been enhancing the diagnostic accuracy of genetic diseases in molecular level as well as paving a new way for basic researches and therapies. 2020 Translational Pediatrics. All rights reserved.

Introduction

The definition of a rare disease is a disease affecting fewer than 2,000 people in Europe (1), while, a disease affecting less than 200,000 people is defined as a rare disease in the United States. Though the chance that individuals who are diagnosed with each rare disease is seemingly low, approximately 7,000–8,000 rare diseases are estimated to date. The Global Genes Project estimates that 300 million people worldwide are affected by a rare disease and eighty percent of rare diseases are gene of origin (2). Moreover, about fifty percent of those affected by rare diseases are in their childhood, thirty percent of who will not survive beyond their fifth year old.

The current genomic technological advancement has changed the research approaches and clinical strategies of rare diseases. In the past few years, the human genome project was firstly completed in mapping all the genes in human with a cost of almost 3 billion dollars. Then the sequencing price significantly dropped with the application of the next generation sequencing (NGS). High throughput and low cost of genomic sequencing make the further insight into genetic diseases in more patients possible. However, the percentage of all known rare diseases with the pathogenic gene is less than fifty percent (3). One reason is that the routine sequencing technology has missed some mutations. Hence, it has still been a challenge to develop diagnostics, managements and genetic advice for these patients in practice.

This paper aims to provide a review of the third generation sequencing (TGS) in genetic diseases. A brief review of revolution of the sequencing technology in genetic diseases is firstly presented. And then, an overview of the TGS approach to the genetic diseases and its clinical effect follows. The bioinformatic methods applied to the new technology and limitations of the TGS are also discussed.

Brief revolution of sequencing technology

Aside from the study of genetic diseases via karyotyping, DNA microarrays, FISH or multiples ligation-dependent probe amplification (MLPA), the first generation sequencing, including Maxam Gilbert methods and Sanger sequencing opened up the new door into the genetic diseases since 1977.

The next generation sequencing (NGS) emerged in 2004 helped researchers gain deeper understanding about the genetic diseases (4). More than 2,400 pathogenic genes have been identified (5) and over 150 genetic diseases have been identified via the whole exome sequencing (WES) (6). Three important centers for Mendelian Genomics (CMGs) funded by NIH, including University of Washington, Yale University and the Baylor College of Medicine utilized the NGS to elucidate many Mendelian disorders (7). In short, the NGS has a great impact on de novo mutations in rare diseases in recent years. However, many rare diseases are still not fully diagnosed by the NGS due to the short-read methods (~150–300 bp). Structural variants (SVs), repetitive elements, extreme guanine-cytosine (GC) content or sequences with multiple homologous elements in the genome are difficult to be characterized via the NGS, even with the use of state-of-the-art bioinformatic algorithms (8,9). These drawbacks of the NGS-based investigations of human diseases have strongly driven the search for other methods to improve the accuracy and reduce diagnosing time in genetic diseases.

The TGS provided by Pacific Biosciences (PacBio) and Oxford Nanopore Technologies (ONT) in 2011, is a single molecular and real-time sequencing technology (10).The PacBio platform adopts single-molecule real-time (SMRT) technology (). In the DNA library preparation, no PCR is required as a closed and circular ssDNA template can be replicated automatically. During the sequencing process, the fluorescence signals are activated by a laser as soon as a labeled dNTPs is incorporated into DNA. A camera system then records the color and duration of the emitted light in real time in the flow cell equipped with zero mode waveguides (ZMVs). The time of the base incorporation is longer as the base is modificated. Thus, the time called “interpulse duration” can indicate the DNA modification event () (11). The SMRT technology also allows the direct RNA-sequencing (12). The SMRT essentially is still based on sequencing by synthesis. Nanopore Sequencing Technology (ONT) utilizes nanopore inserted in an electrical resistant membrane. A potential is applied across the membrane, resulting in a current flowing only through the nanopore (). The characteristic disruptions in the current can be measured, indicating a specific single molecular. In the DNA library preparation, a hairpin structure is designed to ligate the double DNA strands so that the system can read both DNA strands in one continuous read. As dsDNA moves through the nanopore, the bound polymerase or helicase enzyme can attach the DNA in the pore. During sequencing process, a characteristic disruption in the electrical current can be measured as the nucleotide passing through the nanopore () (13). Then the nucleotide can be identified. These features allow the detection of hundreds of kilobases in one continuous read. Ultra-long reads (ULRs) with above 300 kb reads and some close to 1 million bp reads can be sequenced in the ONT (14). Also, the many pocket-sized sequencers developed by the ONT are portable without sophisticated laboratory setup and can be transported out of the lab with low cost. For example, the MinION was transported to Africa for screening Ebola and Lassa virus outbreak (15,16). In short, the features of the TGS introduced by PacBio and ONT allow for the long-read sequencing in real-time with the low alignment and mapping errors during library construction.

Figure 1

SMRT sequencing (image adapted from PACIBO website). SMRT, single-molecule real-time.

Figure 2

A methylated base sequenced by the PacBio; Interpulse duration (dotted arrow in Figure 2A) [image adapted from reference (11)]. PacBio, Pacific Biosciences.

Figure 3

Nanopore sequencing and current signals (Image adapted from Oxford Nanopore Technologies website).

Figure 4

A methylated base (red) sequenced by the ONT. [Image adapted from reference (13)]. ONT, Oxford Nanopore Technologies.

Comprehensive genetic disease identification

The genetic disease can be understood in molecular level rather than that of chromosome owing to the development of sequencing technologies. The TGS has not only helped to discover more novel genetic diseases (17), but also revised the identification of genetic disease (17).

SV has an important role in genetic disorders (18). SVs are defined as mutations affecting more than 50 base pairs. The SVs include deletions, insertions, inversions, mobile element transpositions, translocations, tandem repeats and copy number variants (CNVs) (19). By using SMRT sequencing for two haploid human genomes, Huddleston’s group pointed out that estimated approximately 89% SVs have been missed in the 1,000 Genomes Project (20). Although the sophisticated SV genotyping software methods were available, the detection of SVs was low (30–70%) and the error rates were still high (85%) (21). The single molecular and real-time sequencing has shown a better capacity to discover the structural-variant events. A few SVs related genetic diseases detected through the TGS is reviewed in the following part. For example, Aneichyk and colleagues studied X-linked Dystonia-Parkinsonism (XDP) which is a Mendelian neurodegenerative disease and suggested that a SINE-WNTR-Alu (SVA) mediated aberrant transcriptional mechanism was associated with XDP (22). The precise breakpoints of the deletion in a homozygous 7p14.3 were deciphered in the proband with Barde-Biedl syndrome (BBS) and carrier parents by long-read SMRT sequencing (23). The WES yielded only one heterozygous causal variants in the patient with glycogen storage disease type Ia (GSD-Ia) which is an autosomal recessive disease (24), while, a 7.1 kb deletion covering two exons in G6PC on the other allele were detected through Nanopore long-read whole genome sequencing (WGS) (24). Multiple neoplasia and cardiac myxoma with the negative NGS results were found in connection with a heterozygous 2,184 bp deletion of the first coding exon of PRKAR1A (25). A complex novel translocation t(X;20)(q11.1;p13) was delineated via Nanopore long read sequencing (LRS) technology in a balanced reciprocal translocation (BRT) case (26). Other congenital diseases associated with complex chromothripsis were identified to link to the de novo complex SV breakpoints via ONT (27). In addition, fine-mapping of dipeptidyl-peptidase 6 gene (DPP6) in an autosomal dominant dementia family significantly linked to 7q36 was identified via the PromethION sequencing platform (Oxford Nanopore Technologies) (28).

Another advantage of the TGS is characterize the characterization of complete repeat expansion of genes and discriminate pseudogenes. As a typical example, the C9ofr72'GGGGCC' (G4C2) repeat expansion associated with amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) was validated through Pacific Biosciences and Oxford Nanopore Technologies (29). The CGG short tandem repeats in fragile X syndrome were detected by SMRT sequencing (30). Likely, the familial myoclonic epilepsy was connected with a 4.6 kb repeat expansion and 12.4 kb deletion in complex repeat regions via SMRT (31,32). Additionally, repeat expansions of complex genes, such as ATXN10, HTT, SMAD12, TNRC6A and RAPGEF2 were also validated by SMRT sequencing (33-36). CTG-repeat expansion was confirmed by SMRT in CRISPR/Cas9-mediated editing in myotonic dystrophy patient as well (37). Other complex and challenging regions of the human genome were characterized via the TGS, such as autosomal-dominant polycystic kidney disease (ADPKD). Duplicated and high GC content genomic regions as well as six pseudogenes of PKD1 gene can lead to ambiguous identification of variants via the NGS. However, 94.7% of the patients with PKD1 pathogenic variants were identified via SMAT by Borras and colleagues (38). A GC-rich 60 basepair variable number of tandem repeat (VNTR) and all variants position of the Mucin-1 gene in autosomal dominant tubulointerstitial kidney disease (ADTKD) were also determined by SMRT sequencing (39). Another example is the primary immunodeficiency-associated gene IKBKG. The pseudogene IKBKGP1 can be bypassed by long read, single-molecule sequencing which allows the rapid and efficient identification of the primary immunodeficiency diseases (40). Sanna Gudmundsson and colleagues clarified the mechanism of revertant mosaicism through SMRT (41). They demonstrated that the dominant negative effects of the p.Asp50Asn mutation was reverted by the second-site mutations of Cx25-Asp50Asn resulting in the development of healthy-looking skin in a patient with ichthyosis-deafness (KID) syndrome.

As stated above, the features of the TGS also allows the detection of the epigenetic modification in real time. DNA modifications have been found in a wide range of living organisms, from prokaryotes to eukaryotes. Many existing studies have shown that they play important roles in development diseases, such as lysosomal storage disorders, tumorigenesis, autoinflammatory diseases, imprinting and X chromosome inactivation (42-45). The bisulfite Sanger sequencing and other next generation sequencing have the restriction to read length of only 150–130 bp (46,47). Therefore, long-read single-molecule real-time bisulfite sequencing (SMRT-BS) developed by Yang and colleagues is a technique that combines bisulfite conversion with the TGS and allowed the detection of the targeted CpG methylation in real time (48).

Furthermore, LRS allows the detection of full-length mRNA transcript in one read. The short-read RNA sequencing always leads to inaccurate annotation due to computational transcript reconstruction (49). Aneichyk and colleagues utilized the long-read RNA-sequencing to decipher the TAF1 expression in the X-linked dystonia-parkinsonism (XDP) (22). Roeck and colleagues demonstrated that the Alzheimer’s disease severity was in relation to the varying degrees of nonsense-mediated mRNA decay (NMD) and transcript-modifying events (50). Twenty-seven genetically unsolved patients with an external collagen VI-like dystrophy were found in connection with highly recurrent de novo intronic mutation in COL6A1 via RNA-sequencing (51).

The combination of the TGS with other technologies, such as the NGS or single cell sequencing or target genome editing, will also give an insight into the genome and bring the new therapies (52). Mimori and colleagues utilized the SMRT sequencing and additional short-read data to obtain the high-quality and full-length human leukocyte antigen alleles reconstruction successfully (53). A study of Hendel A and colleagues showed that SMRT sequencing was facilitated to quantify the genome editing outcomes after the large genes were inserted at the endogenous IL2RG, HBB, and CCR5 loci by transcription activator-like effector nucleases (TALENs), zinc finger nucleases (ZFNs) or clustered regularly interspaced short palindromic repeats (CRISPR/Cas9 or RNA-guided endonucleases (RGENs) (54).

Clinical effect of the TGS

Importantly, clinical decisions and outcomes can be benefited from the TGS applications with more complete detection of mutations. For example, de novo mutations can occur in the different stage of embryonic development. Depending on the different stage of postzygotic mutation during development, such mutations may lead to somatic or germline mosaicism or both (55). Understanding of the complex SVs guided the genetic counseling and enable a successful preimplantation genetic diagnosis in the family (56). Maria and colleagues demonstrated that less than 1% of the TCOF1 variant c.3156C>T cells were in the paternal germ cells in a family with a child suffering Treacher Collins syndrome, suggesting the low recurrence risk in offspring (55). Similarly, there were 40% of PTPN11 variant c.923A>C cells in the paternal germ cells in a family with unsuccessful pregnancies, indicating a high recurrence risk of Noonan syndrome in offspring (55). Moreover, AGG interruptions in females with a FMR1 premutation were detected by long-read single-molecule sequencing, which was previously undetected due to the technical difficulties (57). In short, apart from the increasing discovery of novel disease genes, the TGS aids the preimplantation genetic counseling.

Ethics is also important in gene sequencing technology. The informed consent, data privacy and return of results are three issues demanding attention (58). To date, the recommendations of ethical considerations have been addressed by the American College of Medical Genetics and Genomics (ACMG) (59). Obviously, more ethical issues await the TGS as more discoveries of the novel disease genes in clinical practice come up.

With more discoveries of novel SVs, repeat expansions and long noncoding RNAs (IncRNAs) via the TGS, the bioinformatic algorithms have to be TGS-specific and more user-friendly. The major bioinformatic challenges of the TGS is the high sequencing error rate which is 10–15% in the PacBio and 5–20% in the ONT. Therefore, the new alignment and error correction algorithms are required (). Several studies have offered the relatively new methods to correct the sequencing errors in the TGS. The methods for alignment and phasing are LAST (77), BLASR (73), BWA-MEM (74), GraphMap (75), MECAT (64) and minimap2 (78), PBHoney (79), NGMLR (74), Sniffles (74), CORGi (83), SIVM (84), SMRT-SV (95), NextSV (96), NanoSV (97) and Picky (98) in de novo mutations and SVs detection. Regarding the RNA sequencing analysis, the available bioinformatic tools include SQANTI (85), TAPIS (86), ToFU (87), BLAT (88), Gmap (89). In terms of errors correction in sequencing analysis, there are a few available methods as well. Hybrid error correction methods including Nanocorr, MaSCA, PBcR or Spades utilize short-read data to correct the error. However, because of biases in short-read coverage and repetitive sequence, FALCON-sense, HGAP, pbCR, Canu or MARVEL is more accurate as they are self-error correction methods (99). The other technique developed by Jana Ebler and colleagues combined the inference of haplotype and genotypes from noisy long reads (100). A similar software named as NanoSim is a fast and large-scale read simulator to call reads errors in MinION platforms (101). A TGS tool developed by Danze Chen’s group is a bioinformatic suit to compare isoforms, identify alternative splicing pattern and IncRNA (102). Moreover, a time and resource effective strategy for completing short read assembles has been applied, which enable sufficient analysis date to be assembled with the shortest sequencing time (103).

Table 1

Bioinformatic methods in the TGS

Analysis	Methods	Platforms	Applications
Mapping and alignment	MHAP (60)Minimap (61)DALIGNERCanu (62)FALCON (62)Hinge (63)MECAT (64)Miniasm (61)Spades (65)HGAP (66)Flye (67)MARVEL (68)LINKS (69)npScarf (70)RAILLLSPBJelly (71)Ouiver (66)Racon (72)BLASR (73)BWA-MEM (74)GraphMap (75)LASMSA (76)LAST (77)Minimap2 (78)NGMLR (74)PBHoney (79)SMRT-SV (78)Sniffles (58)FALCON-Unizp (80)HapCut2 (81)WhatsHap (82)SIVM (64)NextSV (79)NanoSV (83)Picky (84)	ONT and PacBioONT and PacBioONT and PacBioONT and PacBioPacBioPacBioONT and PacBioONT and PacBioONT and PacBioPacBioPacBioONT and PacBioONT and PacBioONTONT and PacBioPacBioPacBioONT and PacBioPacBioONT and PacBioONT and PacBioONT and PacBioONT and PacBioONT and PacBioONT and PacBioPacBioPacBioONT and PacBioPacBioONT and PacBioONT and PacBioPacBioPacBioONTONT	De novo mutations and SVs detection
	SQANTI (85)TAPIS (86)ToFU (87)BLAT (88)Gmap (89)	ONT and PacBioPacBioPacBioONTPacBio	RNA sequencing analysis
	BaseMods (11)Nanopolish (90)SignalAlign (91)	PacBioONTONT	Methylation analysis
Error correction	Nanocorr (92)MaSuRCA (93)PBcR (94)Spades (65)FALCON-sense (80)Pbdagcon (66)	ONTPacBioPacBioPacBio and ONTPacBioPacBio	De novo mutations and SVs detection

TGS, third generation sequencing; ONT, Oxford Nanopore Technologies; PacBio, Pacific Biosciences; SV, structural variants; SMRT, single-molecule real-time.

The TGS comes with several limitations (14,43,44). First, the DNA library required fresh material or intact cells and the protocols for the handing of ultra-long high molecular weight DNA require improvements. Second, the TGS has the challenges with the higher sequencing error rate and systematic error. Third, the cost of the TGS still has been higher than that of the NGS ($65–$200 per Gb in the PacBio and $22–$90 per Gb in the ONT). Additionally, because the database systems for interpreting complicated SVs are rare, thus the bioinformatic analysis are challenging.

Currently, the NGS is still our first choice of diagnosing the genetic diseases in clinical settings and the TGS can play a complementary role as a result of its limitations. However, with the maturation of the TGS approach, it will be widely used in researches and clinical practice. In the future, the picture of human genome will be more comprehensive as the more genomic data generated.

The article’s supplementary files as