Transcriptomic study of Herpes simplex virus type-1 using full-length sequencing techniques

Herpes simplex virus type-1 (HSV-1) is a human pathogenic member of the Alphaherpesvirinae subfamily of herpesviruses. The HSV-1 genome is a large double-stranded DNA specifying about 85 protein coding genes. The latest surveys have demonstrated that the HSV-1 transcriptome is much more complex than it had been thought before. Here, we provide a long-read sequencing dataset, which was generated by using the RSII and Sequel systems from Pacific Biosciences (PacBio), as well as MinION sequencing system from Oxford Nanopore Technologies (ONT). This dataset contains 39,096 reads of inserts (ROIs) mapped to the HSV-1 genome (X14112) in RSII sequencing, while Sequel sequencing yielded 77,851 ROIs. The MinION cDNA sequencing altogether resulted in 158,653 reads, while the direct RNA-seq produced 16,516 reads. This dataset can be utilized for the identification of novel HSV RNAs and transcripts isoforms, as well as for the comparison of the quality and length of the sequencing reads derived from the currently available long-read sequencing platforms. The various library preparation approaches can also be compared with each other.

Here, we provide a comprehensive overview about preparation methods of sequencing libraries, as well as a description of the data ( Figure 1, Table 3). Detailed statistics about the read quality, such as insertions, deletions, and mismatches, as well as the coverages of the pre-processed data (binary alignment (BAM)) can be found in Table 4 and Fig. 3. The data show that the ONT MinION sequencing resulted in relatively high error rates for insertions, deletions and mismatches. The best read quality (i.e. less insertions, deletions and mismatch)as expectedwas obtained from the PacBio runs. The composition of the errors of the three sequencers (RSII, Sequel, MinION) are different. Mismatches are the most common errors in ONT, while the insertions are the least frequent errors in ONT (consistent with the previously published data 14 ). Intriguingly, contrary to others' datasets 14 , deletions are the major errors in our RSII data; the Sequel shows the same pattern as was expected 14 . The read length distribution of the samples is visualized in Fig. 4 and Fig. 5 Porechop tool (https://github.com/rrwick/Porechop) was used to carry out an analysis on the 5 0 and 3 0 adapters, with which we were able to determine the orientation of the sequencing reads (Table 5 (available online only), Table 6, Fig. 6). About 90% of the PacBio reads could be categorized as forward or reverse oriented read, but only 66%, 58% and 49% from the 1D-seq, Cap-seq and random libraries could be categorized, respectively.
The ratio of the complete (full-length) reads in the different samples were also calculated (Fig. 7). The data shows that the random-primed cDNA sequencing as well as the dRNA sequencing resulted in very few complete reads, which has a technical reason: random primed sequencing could gain complete reads when the random primer binds to the 3 0 -end of the RNA. The small number of full-length reads, from the dRNA sequencing is consistent with our previous findings: we observed that the dRNA sequencing resulted in poor 5′ and 3′ read ends 15,16 . It can also be seen that nanopore sequencing produces a large number of incomplete reads; these short fragments are probably removed from PacBio sequencing by the MagBead loading.

Methods
A part of the methods section (PacBio RSII sequencing) is an expanded version of descriptions in our related work 7 , however, the largest part of the data including the entire PacBio Sequel and ONT sequencing data has not yet been published elsewhere. The various library preparation and sequencing methods utilized in this study are shown in Figure 1.

Viruses, cells and infection
The Vero immortalized kidney epithelial cell line, derived from the kidney of an African green monkey (Cercopithecus aethiops) was used for maintaining and propagating HSV-1. The cell culture was grown in Dulbecco's modified Eagle medium (DMEM, Gibco/Thermo Fisher Scientific) with 10% Foetal Bovine Serum (Gibco/Thermo Fisher Scientific) and 100 μl/ml Penicillin-Streptomycin 10 K/10 K Mixture   (Lonza), and in a 37°C incubator with a humidified atmosphere of 5% CO 2 in air. The viral stocks were prepared by infecting rapidly-growing semi-confluent cells with viruses at a multiplicity of infection (MOI) of 1 plaque-forming unit (pfu)/cell. The infected cells were incubated until complete cytopathic effect was observed. Three times freeze-thaw cycles were applied and it was followed by a centrifugation step at 10,000 g for 15 min. Cells were infected in a suspension with HSV-1 at an MOI of 1, then they were incubated for 1 h. The virus suspension was removed and cells were washed with phosphatebuffered saline (PBS). This step was followed by the addition of fresh medium to the cells and they were incubated for 1, 2, 4, 6, 8, or 12 h for the RSII sequencing and for 1, 2, 3, 4, 5, 6, 8, 12, 18, or 24 h for the Sequel and MinION runs.

RNA purification
Total RNA extraction. Total RNA isolation was carried out by using the NucleoSpin® RNA kit (Macherey-Nagel) following the kit's recommendations and our previously published methods 12 . In sum, the viral infected cells were lysed in a lysis puffer (containing chaotropic ions which inhibit ribonucleases, supplied by the kit). DNase I (provided by the kit) treatment was carried out during the purification. The RNA samples, which were bound to a silica membrane, were eluted in RNase-free water. The potential residual DNA contamination was eliminated by applying additional DNase treatment; the Ambion® TURBO DNA-free TM Kit was used (Thermo Fisher Scientific). The final concentrations of the RNA samples were measured by Qubit® 2.0 Fluorometer using Qubit RNA BR Assay Kit (Life Technologies) and then they were stored at −80°C until further use. The RNA samples taken from each time points were mixed for library preparation and sequencing.
Ribosomal RNA depletion. For the random hexamer-primed cDNA sequencing, the RNA samples were handled with the ribodepletion kit (Epicentre Ribo-Zero TM Magnetic Kit H/M/R, Epicentre/ Illumina), to remove the ribosomal (r)RNAs.

Preparation of cDNAs and sequencing libraries
PacBio SMRTbell library preparation. Full-length cDNAs were prepared according to the PacBio Isoform Sequencing (Iso-Seq) protocol by using the Clontech SMARTer PCR cDNA Synthesis Kit. We applied the no Size Selection protocol for the study of short RNAs, while we carried out size selection by using SageELF TM and BluePippin TM Size-Selection Systems (Sage Science) for the isolation of long transcripts. Reverse transcription (RT) reactions were primed with oligo(d)T (supplied by the Clontech Kit) or random hexamer primer (custom designed, ordered from IDT DNA). Samples were amplified by PCR using KAPA HiFi Enzyme (Kapa Biosystems), according to the above mentioned PacBio protocol. In short, primary denaturation was done at 95°C for 2 min. This step was followed by 16 cycles for PAsequencing, as well as 20 or 30 cycles for random hexamer-primed samples (the optimal cycle was determined in the optimization step) at 98°C for 20 s (denaturation), 65°C for 15 s (annealing) 72°C for 4 min (extension). 72°C was used for the final extension, it was set to 5 min. (n: 18 cycles was set to the No size-selection protocol, and the same PCR setting was applied for size-selected samples.     PCR products were mixed together and then size selected with the SageELF TM System following the PacBio's protocol. Size-selected cDNAs were amplified with KAPA enzyme using the above mentioned conditions. The cDNAs with a size over 5 kb was size-selected again, with the BluePippin TM System to remove the short samples. Five-hundred ng of each non-size-selected sample was used for the generation of the SMRTbell templates, with the PacBio DNA Template Prep Kit 1.0. The quantity of the size-selected cDNAs used in the SMRTbell template preparation reaction was based on the following protocols: Procedure & Checklist -Isoform Sequencing (Iso-Seq TM ) using the Clontech SMARTer PCR cDNA Synthesis Kit and SageELF TM Size Selection System & BluePippin TM Size-Selection Systems. The DNA/ Polymerase Binding Kit P6 and v2 primers were used to generate SMRTbell library-DNA polymerase complexes. The ready complexes were bound to MagBeads using the PacBio MagBead Binding Kit.
The volume of the sequencing primer for the annealing, and the polymerase (P5 or P6) for the binding was determined using the PacBio Calculator version 2.3.1.1. by adding the concentrations and the average insert sizes of SMRTbell templates.
The polymerase-template complexes were bound to MagBeads, loaded onto SMRT Cells and sequenced on the RSII instrument. of the reads mapped to the HSV-1 genome are shorter than 3,000 bps, most of the reads are between 1,001-2,000 bps, independently from the focus of the size selection. (d) This bar chart shows that the same library preparation method (Isoform sequencing with no size selection using Clontech SMARTer PCR cDNA Synthesis Kit) result in different read length distribution using the two PacBio platforms: 40% of the Sequel reads are longer than 2,000 bp, while only 10% of the RSII reads belong to this size range. (e) The average read length using the MinION nanopore sequencing is below 2,000 bp. There is no random primed read above 9,000 bp. Intriguingly, the longest reads are derived from the direct RNA sequencing approach (15,000-19,000 bp).  This bar chart compares the data derived from the two PacBio approaches. We could detect the adapters on more than 90% Sequel reads, while we obtained in a poorer result from RSII. (c) This diagram compares the ONT's own library preparation approach (1D) with the Lexogen Teloprime Cap-selection protocol, which used for a library, run on a MinION flow cell. (d) We were not able to identify the orientation of about 50% of reads using the random-primed approach, while dRNA-seq could detect adapters on about 20% of the reads. (e) Segmented pie chart illustrates the distribution of reads with (labelled as: Fw and Rev) or without orientation (labelled as: No orientation). Fifty-six % of this dataset derived from the different nanopore sequencing approaches, 26.5% from Sequel and the remaining less than 20% is from the RSII. A high ratio (~90%) of reads from PacBio RSII and Sequel sequencing could be categorized as Fw or Rev, however Porechop were not able to identify the adapters and thus the orientation of at least about 30% of reads.  The Binding Calculator (PacBio) was used to calculate the amount of samples to be used for sequencing, following MagBead one-cell per well (OCPW) method. The P6v2 binding kit was used, the on-plate concentration was set to 0.05 nM. The insert sizes were set based on the applied size-selections: 1000, 2500 and 6000 bp sizes were chosen.
Briefly, the sequencing primer was diluted in PacBio Elution Buffer (EB) to 150 nM. The annealing step was carried out using 1 μl template (cc:~20 ng/μl), the diluted sequencing primer and primer buffer (10×). The concentration of this reaction mixture was 0.8333 nM. Annealing step was set to 20°C for 30 min. The DNA polymerase enzyme was diluted to a final concentration of 50 nM in PacBio Binding Buffer (BB) v2, and then it was bound to the annealed sample followed by the addition of DTT, dNTP and BB. The sample complex (final concentration was set to 0.5 nM) was incubated at 30°C for 4 h. The complex (0.5 μl) was added to 18.5 μl MagBead BB (final concentration was set to 0.0125 nM). The detailed protocol of the MagBeads preparation is as follows: 73.9 μl MagBeads were washed with MagBead Wash Buffer, then it was replaced by MagBead BB (73.9 μl). The washed MagBead was used to bound to the sample-complex. Nineteen μl from the complex was added to the washed MagBeads and then they were incubated in a HulaMixer (Life Technologies) at 4°C for 30 min. After this binding step, the sample was purified with BB (19 μl), then with 19 μl WB. The final elution was in 19 μl BB. The MagBead-bound sample complexes were loaded for sequencing.
The MagBead One Cell Per Well protocol was applied for the sequencing. Sequencing runs were performed by using the PacBio RS II sequencer or Sequel platform. Size-selected and no size-selected samples were run on RSII, while the Sequel platform was applied for the no size-selected samples. The DNA Sequencing Reagent 4.0 was used for the reactions. Two hundred and forty min or 360 min movie lengths were set for the RSII runs, while 600 min was applied for the Sequel sequencing (one movie was recorded for each SMRT Cell).

Direct RNA sequencing
To avoid the potential false-priming effect of PCR reactions, an amplification free method was used for the analysis of HSV-1 transcripts. For this, the ONT's Direct RNA sequencing (DRS) protocol (Version: DRS_9026_v1_revM_15Dec2016) was utilized. A total RNA sample containing an equal amount from the 10 time points (1,2,3,4,5,6,8,12,18,24 h pi) was used to isolate the polyadenylated RNA fraction. This sample (115 ng) was then used as template for the RT, it was added to the RT adapter (supplied by the DRS kit) ( Table 7). T4 DNA ligase (2 M U/ml; New England BioLabs) was also added to the reaction. The reaction mixture was incubated at room temperature for 10 min. SuperScript (SS)III RT enzyme (Life Technologies) was added to the RNA-cDNA hybrid, and the generation of the first-strand cDNA was carried out according to the ONT DRS kit's recommendations: the reaction was done at 50°C for 50 min, then the SSIII was inactivated at 70°C (10 min). The RNase OUT (40 U/μl; Life Technologies) recombinant RNase inhibitor-treated Agencourt AMPure XP Magnetic Beads (Beckman Coulter) was used to purify the sample (2U RNase OUT per 1 μl bead), then it was eluted in 20 μl Nuclease-Free Water (Ambion/Thermo Fisher Scientific). T4 DNA ligase and NEBNext Quick Ligation Reaction Buffer (New England BiceoLabs) were used to ligate the purified sample to the RMX adapter. The ligation was done at room temperature. The incubation time was 10 min. The adapter-ligated sample was purified with the RNase OUT-handled XP beads using Wash Buffer (DRS Kit), and then eluted in 21 μl Elution Buffer (DRS Kit). The quantification of the ready libraries was performed using the Qubit 2.0 Fluorometer as well as Qubit dsDNA HS Assay Kit (both from Life Technologies). The R9.4 SpotON Flow Cell was used for MinION sequencing.
Oxford Nanopore 1D cDNA sequencing Samples were sequenced on a MinION device from the ONT according to the ONT 1D Strand switching cDNA by ligation protocol (Version: SSE_9011_v108_revS_18Oct2016). The sequencing libraries were  generated by using the ONT Ligation Sequencing Kit 1D (SQK-LSK108). The PolyA(+)-selected RNA fraction or the rRNA-depleted sample was used for cDNA production. Identical quantity of RNA samples from each of the applied time points were mixed together. Thirty-one ng from the Poly(A + )-selected, or 50 ng from the rRNA-depleted sample was used as template for the RT reactions. An anchored oligo(d)T primer [(VN)T20; ordered from Bio Basic, Canada, (Table 7)] or a modified random-primer (Table 7) was used for priming the reactions. The RNA-primer mixture was incubated at 65°C for 5 min. After this short step, a strand-switching oligo [containing three O-methyl-guanine RNA bases (PCR_Sw_mod_3G; Bio Basic, Canada)], buffer and DTT [both are derived from the SuperScript IV Reverse Transcriptase kit (Life Technologies)] were added to the reactions. The samples were handled with a recombinant RNase inhibitor (RNase OUT TM , Life Technologies); the incubation was carried out at 42°C 2 min.
The SuperScript IV Reverse Transcriptase enzyme (200 unit) was added to the sample. The RT reaction was carried out in a Veriti Cycler (Applied Biosystems) at 50°C for 10 min. It was followed by the strand-switching step at 42°C for 10 min. Samples were heated to 80°C (10 min) to inactivate the enzymes. Samples were amplified by PCR with KAPA HiFi DNA Polymerase (Kapa Biosystems) and Ligation Sequencing Kit Primer Mix (1D Kit). Five μl from the cDNAs were used as template in each of the PCR reactions. The Veriti PCR machine was set as follows: initial denaturation 95°C, 30 sec (1 cycle); denaturation 95°C, 15 sec (15 cycles); annealing 62°C 15 sec (15 cycles); extension 65°C 4 min (15 cycles); final extension step 65°C, 10 min. NEBNext End repair/dA-tailing Module (New England Biolabs) was utilized for repairing the DNA ends, while the NEB Blunt/TA Ligase Master Mix (New England Biolabs) was used for ligating the adapters (1D kit). The Beckman Coulter Agencourt AMPure XP beads were used to purify the DNA after each of the enzymatic steps. The concentration of the samples was measured by using the Qubit Fluorometer 2.0 (Life Technologies) and the Qubit (ds) DNA HS Assay Kit (Life Technologies). The libraries were sequenced on the ONT R9.4 SpotON Flow Cells.

MinION cDNA sequencing on Cap-selected samples
To generate full-length cDNA from capped and polyadenylated RNAs, the "all-in-one" protocol from the Lexogen was applied using the TeloPrime Full-Length cDNA Amplification Kit. A 2.2 μg mixture from the different total RNAs (1,2,3,4,5,6,8,12,18 and 24 h pi) was used for the RT. For this, the sample was mixed with RT buffer and a specific primer (both are the part of the kit, Table 7). The reaction started at a 30 sec incubation at 70°C, then it was followed by a 1 min step at 37°C. At this point, the reaction was kept at 37°C. The reverse transcriptase and the additional reagents (derived from the kit) were mixed with the sample and then the incubation was containing at 37°C for 2 min. The next step of the RT reaction was carried out at 46°C for 50 min. Sample was purified by using the kit's Silica columns. The double-strand (ds) specific ligase enzyme (Lexogen kit) was used to join the adapter to the cDNA, the reaction was performed at 25°C, overnight, then the sample was purified using the silica membranes of the Lexogen kit. The dscDNAs were generated by using the Enzyme Mix and the Second-Strand Mix (Lexogen kit). The cDNA production was performed in a Veriti cycler. The following protocol was applied: 98°C for 90 sec, 62°C for 60 sec, 72°C for 5 min (16 cycles   concentration was detected by using Qubit 2.0 and Qubit dsDNA HS quantitation assay (Life Technologies). The specificity of the obtained product was checked by using real-time PCR. The Rotor-Gene Q qPCR machine (Qiagen), a gene specific primer (us9, 10 μM each, ordered from IDT DNA, Table 8), and the ABsolute qPCR SYBR Green Mix (Thermo Fisher Scientific) was applied. The preliminary denaturation was carried out at 94°C for 15 min, then 35 cycles of 94°C for 25 sec, 60°C 25 sec and 72°C 6 sec was applied. The cDNAs from polyadenylated-capped RNA samples were used for library preparation for ONT sequencing. The 1D Strand switching cDNA by ligation method was used. After the end-repair, the samples were ligated to the 1D adapters. Finally, they were measured on the ONT R9.4 SpotON Flow Cells.

Read processing
The SMRT Analysis v2.3.0 (PacBio RSII), the SMRT Link (PacBio Sequel) and the Albacore v2.0.1 (ONT MinION) software packages were used for base calling (Figure 1). The reads were mapped by using GMAP and the following setting was applied: Minimum Full Passes = 1, Minimum Predicted Accuracy = 90, Minimum Length of Reads of Insert = 1, Maximum Length of Reads of Insert = No Limit. These consensus reads were mapped using GMAP 17 , with the following settings: gmap -d Genome.fa --nofails -f samse File.fastq > Mapped_file.sam. The quality information was acquired by using custom made routines 18 .
The Porechop v.0.2.3 software was used to determine the orientation of the sequencing reads. For this, a modified adapter.py file was used, where we added the various, library-specific adapter sequences ( Table 6). The first mapped nucleotide downstream the "END" adapter was labelled the 5 0 end of a read, while the last mapping nucleotide upstream of the "START" (polyA tail: A 20 or 3 0 adapters) was designated the 3 0 end of the read. Reads lacking an adapter on both ends, or with 5 0 or 3 0 adapters on both ends, can be discarded from further in silico analysis.

Data Records
Data from MinION and Sequel sequencing have been uploaded to the European Nucleotide Archive (Data Citation 1) -contains BAM files. The RSII raw sequencing files, processed data files as well as metadata have been submitted to the Gene Expression Omnibus repository (Data Citation 2). All reads were mapped to the X14112 genome build. The provided sequencing data can be used without restrictions.

Technical Validation
The quantity of the purified total RNAs, the polyA-selected RNAs, the ribodepleted RNA samples, as well as the cDNA samples and the final sequencing libraries were detected by Qubit 2.0 (Life Technologies) fluorometer using the Qubit RNA Broad-Range, High Sensitivity RNA and High Sensitivity dsDNA Assay Kits.

Usage Notes
Our provided dataset was primarily generated to analyse the potential splice variants, as well as transcriptional start and stop site variations, the isoforms and the complexity of HSV-1 transcriptome. The provided raw RSII data files can be utilized to develop novel base calling algorithms or read processing tools, as well as to improve the currently existing bioinformatics software. The uploaded BAM files contain reads already aligned to the X14112 HSV-1 reference genome using GMAP v2017-04-24 17 .