Ploidy-stratified single cardiomyocyte transcriptomics map Zinc Finger E-Box Binding Homeobox 1 to underly cardiomyocyte proliferation before birth

Mice

C57BL/6 J mice were obtained from Taconic Europe, housed with a 12/12 h light/dark cycle, and fed ad libitum. For scRNA-seq, mice were plug bred, and litters for each of the three timepoints were obtained from different breeding pairs. Plug was checked in the morning and evening. For E16.5 primary cultures, mice were plug bred as well, whereas for primary P0 cultures continuous breeding was used. All animal experiments were approved by the Danish Council for Supervision with Experimental Animals (#2016–15-0201–00,941 and #2022–15-0201–01119).

Preparation of CMs for scRNA-seq

For scRNA-seq, the left heart ventricle was dissected under a stereomicroscope at E16.5, P1, and P5 (n = 3 litters, each counting 4–8 pups), and enzymatically dissociated using the semiautomatic GentleMACS tissue dissociator system (MACS Miltenyi Biotec; Neonatal Heart Dissociation Kit 130–098-373) according to the manufacturer’s recommendations. Following viable cell counting (NC-200, ChemoMetec), dissociated cells were stained with a fixable viability stain (Fixable viability stain 570; BD Biosciences, 564,995) prior to fixing in methanol for 15 min followed by rehydration to reverse the RNA to its original state. During rehydration, the RNase inhibitor, RNasin Plus (Promega; N2615), was added to prevent RNA degradation and included in all subsequent steps. After rehydration, samples were stored at -80 °C until analysis. All reagents were high grade, RNase free and the environment was kept strictly RNase free to avoid degradation of the RNA. For comparison of fresh and fixed scRNA-seq profiles, mouse myoblasts (C2C12; ATCC, CRL-1772) were used and maintained as recommended.

Fluorescence-activated cell sorting (FACS)

Fixed cardiac cells were stained for the CM marker MYH1 (Mouse IgG2b,k; 1:300; MF20-c; DSHB) and visualized by donkey anti-mouse IgG Alexa Fluor 488 (1:200; Invitrogen, A21202), whereas Hoechst 33342 (Sigma) was added 5 min before sorting (FACSAriaIII, BD Biosciences). Prior to FACS, cells were filtered (Falcon, 352235) to avoid cell clumps. Strict RNase free conditions as described above including new tubing were prioritized throughout the procedure. Analysis and sorting gating strategy (Supplementary Fig. 1c) included hierarchical gating using the FACSDiva software v8.0.1 (BD Biosciences) based on FSC/SSC, viability Alexa 570, and MYH1-Alexa 488, and Hoechst 33342. For each developmental stage (E16.5, P1, P5), three independent sortings (n = 3, each consisting of cells from one litter) were performed after carefully checking and validating the FACS setup using FMO controls (Supplementary Fig. 1c). Prior to scRNA-seq analysis CM purity, nuclei number, and cell clumping of sorted cells were assessed using immunofluorescence microscopy whereas the RNA integrity number (RIN) was determined using an Agilent 2100 Bioanalyzer (Agilent Technologies) combined with the Agilent RNA 6000 Nano Kit (Agilent Technologies) [56]. Sorted cells were stored at -80 °C until scRNA-seq.

ScRNA-seq

For scRNA-seq, cells originating from three independent FACS were pooled (14–20 pups/sample) to account for biological diversity in the scRNA-seq analysis. Single Cell 3’ RNA-Seq libraries were prepared using Chromium Single Cell 3′ Reagent Kits v2 (10 × Genomics) according to the user guide. In brief, cellular suspensions of approx. 1200 cells/µl were mixed with master mix reagents and loaded on a Single Cell A Chip (10 × Genomics) together with Single Cell 3’ Gel Beads (10 × Genomics) and partitioning oil to generate single cell gel beads-in-emulsion (GEMs). The GEM generation took place in a Chromium Controller (10 × Genomics). Single cell reverse transcription was performed in a standard thermal cycler, and the GEMs were subsequently broken using Recovery Agent (10 × Genomics). The resulting cDNA was cleaned up with DynaBeads MyOne Silane Beads (Thermo Fisher Scientific) and SPRIselect Reagent Beads (Beckman Coulter), and then amplified by PCR using Single Cell 3′ Reagent Kit v2 (No. of cycles: 8). After another cDNA clean-up with SPRIselect Beads, the fragment sizes and concentrations were measured using QIAxcel DNA High Resolution Kit (1200) (Qiagen) and Qubit dsDNA HS Assay Kit (Thermo Fisher Scientific), respectively. Enzymatic fragmentation, end-repair, and A-tailing were performed in one-step using the Single Cell 3′ Reagent Kit v2, and fragments of approx. 200 bp were selected by double sided-size selection using SPRIselect Beads. NGS libraries were then constructed by adapter ligation and PCR mediated sample indexing (No. of cycles: 13). After a final double-sided size selection, the NGS library concentrations were measured using Qubit dsDNA Assay Kit. Libraries were sequenced on the Illumina NextSeq 500 platform using NextSeq 500/550, high output Reagent Cartridge V2, Illumina Kit (Read 1 = 26 cycles, i7 Index = 8 cycles, Read 2 = 130 cycles), and the second analysis was performed on a Illumina NovaSeq 6000.

ScRNA-seq data analysis

Read alignment and construction of gene expression matrix Base calls were converted to FASTQ format and demultiplexed using the cellranger mkfastq function embedded in the 10 × Genomics cellranger software package using default settings (https://support.10xgenomics.com/single-cell-gene-expression/software/overview/welcome). Single cell gene counts matrices were generated using the cellranger count command. During this step, FASTQ files generated by the cellranger mkfastq step, were aligned to the Mus musculus genome (mm10/GRCm38) using the splice-aware aligner STAR [12]. Subsequently, STAR used the Mus musculus transcriptome reference (GRCm38.84) to segregate the mapped reads into exonic, intronic and intergenic regions and for assessment of how confidently the reads have been mapped to these regions. Only non-duplicated reads which were confidently mapped to the transcriptome, and which had barcodes and unique molecular identifiers (UMIs) were used for UMI counting. The expression matrices were generated by counting the number of strand-specific UMI for each cell mapping to either the exonic or intronic regions of each gene.

Clustering and UMAP visualization Using the R package Seurat[62] v 2.3.0 and 3.1.5 dimensionality reduction by principal component analysis (PCA) was performed; subsequently the PCA data analysis was used as input for visualization by Uniform Manifold Approximation and Projection (UMAP) clustering [69, 70]. Cell clustering by expression pattern was performed by first calculating the k-nearest neighbors and constructing the shared nearest neighbor (SNN) and next optimizing the modularity function to determine clusters.

Clustering and Heatmaps The Seurat v 2.3.0 and 3.1.5 [7, 62] were used for cluster visualization by UMAP and for differential gene expression of marker genes between clusters.

For each of the samples, a Seurat object was created, and the cells filtered based on whether they expressed a combination of the CM-specific markers Tnni3, Tnnt2, Actc1, and Tnnc1. Each sample was then log-normalized, variable features were identified using “vst” as selection method and 2000 nfeatures, and the data was scaled using nCount_RNA for vars.to.regress. PCA was run and based on jackstraw- and PC elbow plots the optimal number of dimensions was determined (range 9–14). Moreover, when applied, all samples were merged, integrated using FindIntegrationAnchors and IntegrateData, filtered, normalized, and scaled as described above with generated UMAP plots depicting cell cycle phases, clusters, and original data affiliation as for each individual sample.

Visualization, clustering, and cell cycle analysis UMAP plots and clusters were generated as described above using PCA as reduction type and resolution = 0.6; based on the top 30 marker genes for each cluster. Subsequent GO term enrichment was evaluated using clusterProfiler::enrichGO [75, 78] and the “org.Mm.eg.db” library with ont = “BP”, pAdjustMethod = “BH”, and cutoff values = 0.01. Features witg avg_log2FC > 0.5 were used, where each cluster was named according to biological identity. Finally, each dataset was split into three groups (G1-, S-, or G2/M-phase) based on the expression of cell cycle markers [67] and each cell was assigned with a cell cycle score using Seurat::CellCycleScoring.

Analysis across developmental stages After merging and integration as described above, two clusters of cell cycle active E16.5 and P5 cells, respectively, were subtracted from the data and compared using FindMarkers. The resulting list of features was used for generating cnetplot and TF analysis. Mouse single site analysis was used for TFs (oPOSSUM version 3.0) [25, 26, 33], all genes in current dataset as background, all vertebrate profiles with a minimum specificity of 8 bits, conservation cutoff 0.40, matrix score threshold 85%, up/downstream sequence 5000/5000). In addition, oPOSSUM results were supported by GSEA using the Molecular Signature Database [41, 42, 63] and transcription factor targets (TFT).

Trajectory analysis Data were prepared in Seurat (filtered and cell cycle assigned; since UMI data were used, normalization was avoided in agreement with recommendations by the Monocle platform) in merged pools of either 2n-, 4n-, or all samples, respectively. Subsequently, phenotype data and feature data were extracted from the Seurat object and converted to a Monocle CellDataSet (CDS) object. Next, dispersion estimates for count dataset were obtained using monocle::estimateDispersions and cells were sorted according to num_genes_expressed (500 < num_genes_expressed < 3000). A set of ordering genes was isolated using differentialGeneTest and used to order the CDS by the monocle::setOderingFilter. Next, the dimensions were reduced and cells were ordered along the trajectory using monocle::reduceDimension and monocle::orderCells, respectively. The trajectory was plotted depicting original identity, cell cycle phase, and pseudotime state. The monocle::BEAM function was utilized in each branch point of the trajectory plots to evaluate branch point dependent gene expression.

Plasmids and AAV9 packaging

To determine the most efficient AAV serotype for CM transduction, pilot studies with both AAV6 and AAV9 transduction were performed, as these serotypes have previously shown efficient in CM transduction [53]. In our study design we found the AAV9 serotype to be much more efficient in transducing CMs, as compared to the AAV6 serotype (data not shown).

Generation of plasmids Plasmids harboring the genes of interest were purchased from Origene (Mouse Tagged ORF Clones; Supplementary Table 1), except for Nfya, which were synthesized by GeneArt (Thermo Fisher; Supplementary Table 1). The AAV backbone transfer vector was derived through modifications of the plasmid pAAV-EF1a-mCherry-IRES-Cre (a gift from Karl Deisseroth; Addgene plasmid # 55,632; http://n2t.net/addgene:55632; RRID:Addgene_55632) [15], allowing simultaneous transcription of mCherry and the gene of interest through the internal ribosome entry site (IRES). Thus, due to the IRES site, transcription of the gene of interest correlates to the level of mCherry. To unify the process of gene insertions, the restriction sites SgfI and MluI were inserted into the plasmid. Briefly, the already existing MluI restriction site was removed by introducing a point mutation in the plasmid by PCR amplification using the following primers: Forward: CGCACGGGTAAGCTTTGCAAAGATGGATAAAGTTTTAAACAGAGAGGA and Reverse: AAGCTTACCCGTGCGGCCGCAGGAACCCCTAGTGAT. The Cre site was then removed and the SgfI and MluI restriction sites were hereafter inserted by PCR amplification using the primers Forward: TCTGGTGCGATCGCCTAGACGCGTTAGATTCGATATCAAGCTTATCGATAATCAACCTCT and Reverse: CTAGGCGATCGCACCAGAACCACCATTATCATCGTGTTTTTCAAAGGAAAACCACGTCCC. Finally, a truncated chicken cardiac Troponin T promoter [53] (cTnT promoter; synthesized by GeneArt (Thermo Fisher); the DNA sequence was kindly provided by Professor Brent A. French, University of Virginia, USA) was inserted for CM specificity by PCR amplification in two steps: first, plasmid pAAV-EF1a-mCherry-IRES was PCR amplified by primers (Forward: GGAATTCCATATGGGTACCGGATCCGTGAGC and Reverse: GCTCTAGAAATTCCCACTCCTTTCAAGACCTAG) containing the XbaI and NdeI restriction sites to excise the EF1a promoter. Secondly, the cTnT promoter was inserted between the two restriction sites (Forward: GCTCTAGAGCAGTCTG and Reverse: GGAATTCCATATGAGGTC). The resulting pAcTnT-mCherry-IRES plasmid was then sequenced (Eurofins Genomics, Ebersberg, Germany) for validation (Data not shown). Genes (Origene plasmids and Nfya) were inserted into the pAcTnT-mCherry-IRES plasmid between the SgfI and MluI restriction sites. Since the SgfI restriction site was already included in the Egr1 sequence, Egr1 was amplified by the following primers Forward: AATGGTGGTTCTGGTGCGATCGCATGGCAGCGGCCAAG and Reverse: TTGATATCGAATCTAACGCGTGCAAATTTCAATTGTC. Next the Egr1sequence was added to pAcTnT-mCherry-IRES by NEBuilder® HiFi DNA Assembly Master Mix (NEB). Proper gene insertions were validated by enzymatic digestion at the respective restriction sites and size determined by gel electrophoresis (Data not shown).

For plasmid packaging in an AAV9 serotype capsid we used the Rep/Cap plasmid, pAAV2/9n, a gift from James M. Wilson (Addgene plasmid # 112,865; http://n2t.net/addgene:112865; RRID:Addgene_112865) and the helper plasmid pHelper (a kind gift to our collaborator Per Svenningsen, University of Southern Denmark, from Ben Deverman, Caltech, Pasadena, USA).

Virus generation Large-scale AAV generation for in vitro use was performed in HEK293T cells (ATCC; CRL-3216) by co-transfection with pAcTnT-mCherry-IRES (empty vector) or pAcTnT-mCherry-IRES harboring the gene of interest, pAAV2/9n and pHelper. Transfection efficiency was addressed by mCherry visualization using immunofluorescence microscopy. Five days after transfection, recombinant AAV was isolated by PEG 8000 precipitation and purified by iodixanol gradient ultracentrifugation followed by centrifugation through an Amicon Ultra Centrifugal filter (50 K). Recombinant AAV yields were determined by quantitative real-time PCR (qRT-PCR) through a titration of pAcTnT-mCherry-IRES plasmid using the primers Forward: AGTGTTGCATTCCTCTCTGG and Reverse: AGCGCATGAACTCCTTGAT.

Adenoviral constructs were generated by Vector Biolabs (PA, USA) using Adenoviral Human Type 5 (dE1/E3) as backbone. For ZEB1 knockdown experiments, a U6 promoter was driving ZEB1 short-hairpin RNA (shRNA) expression of the sequence 5´CCGGATAGAGGCTACAAGCGCTTTA-CTCGAG-TAAAGCGCTTGTAGCCTCTA-TTTTTTG-3´ and a targeting sequence of ATAGAGGCTACAAGCGCTTTA. An eGFP reporter was expressed under a separate CMV promoter. Ad-GFP-U6-scrmb-shRNA (cat. no. 1122N) containing a scrambled shRNA and an eGFP reporter was used as control. For ZEB1 overexpression experiments, the backbone vector contained a CMV promoter to drive expression of the gene of interest. Ad-GFP-Zeb1 was generated using mouse cDNA (GenBank: BC139768.1) and eGFP, and ZEB1 were expressed under separate CMV promoters. Ad-GFP (cat.no. 1060) was used as empty control.

E16.5 CM cultures and Zeb1 knockdown

On embryonic day 16.5 (E16.5), the pregnant mice were sacrificed by cervical dislocation and the hearts from the pups were quickly removed and placed in a cardioplegic buffer (MIB; 1.2 mM KH2PO4 (pH 7.4); 0.25 g/l Na2CO3; 6.44 g/l NaCl; 2.6 mM KCl; 1.2 mM Mg2SO4; 11 mM glucose) supplemented with 1% Bovine Serum Albumin (BSA; MIB/1%BSA). The heart ventricles were dissected under a stereomicroscope before enzymatically dissociation into a single cell suspension using the semiautomatic GentleMACS tissue dissociator system as described by the manufacturer. Dissociated cells were counted (NC-200; ChemoMetec), plated on extracellular matrix (ECM) at a density of approx. 118,500 cells/cm2, and cultured in growth medium (79.5% DMEM (supplemented with 1% PenStrep (PS)), 19.5% Medium 199 (supplemented with 1% PS), and 1% newborn calf serum). After 24 h, the number of cells were counted in some wells to calculate the amount of virus required. Optimal MOI was determined from titrating the virus and quantifying transduction efficiency as well as observing for immediate cytotoxicity.

After 24 h of culturing, cells were transduced with 10 MOI of either Ad-GFP-shRNA or Ad-GFP-shRNA-Zeb1. In addition, 10 µM of 5-ethynyl-2′-deoxyuridine (EdU) was added to assess for cell cycle activity. The medium, with or without EdU, was replenished every 24 h, and experiments were terminated as indicated at 96 h after transduction for analysis.

Neonatal CM cultures and viral transduction

Neonatal (P0) mouse pups from each litter were sacrificed by decapitation, whereafter the hearts were quickly removed, and the ventricles dissected under a stereomicroscope. Dissected ventricles were pooled in a tube with MIB/1%BSA before enzymatic dissociation into a single cell suspension using the semiautomatic GentleMACS tissue dissociator system as described by the manufacturer. Dissociated cells were resuspended in growth medium and the number of cells were counted (NC-200; ChemoMetec). Cells were seeded in 12-well plates pre-coated with ECM at a density of approx. 236,500 cells/cm2 and placed in an incubator (37 °C, 5% CO2) or approx. 88,235 cells/cm2 on 4-well chamber slides (cat.no. 154917, Lab-Tek™ II) for confocal microscopy.

Before deciding which concentration of virus to use, both for AAV9 and adenovirus transduction, titration tests were performed and the concentrations resulting in the most efficient transduction without causing cytotoxicity were used. After 24 h the number of cells in each experiment were estimated (NC-200; ChemoMetec), and cell cultures were transduced with either 750,000 viral genomes (vg)/cell of the desired AAV9 or 50 MOI of adenovirus. For AAV9 experiments, six, 24, and 48 h after viral transduction, medium was refreshed with medium containing 10 µM EdU. For adenovirus, EdU was added together with the virus and replenished every 24 h. All cells for qRT-PCR were replenished with medium without EdU. Cells were either fixed in 2.5% Neutral Buffered Formalin (NBF) diluted in HBSS/5%FBS/1%PS for flow cytometry analysis 72 h post transduction (see below), fixed in the wells in 10% NFB or 4% Paraformaldehyde (PFA), or the RNA was isolated for qRT-PCR 48 h post transduction for adenovirus or 72 h post transduction for AAV9 (see below). Transduction efficiency was addressed by immunofluorescence microscopy for mCherry during culture, and by qRT-PCR (mCherry and/or gene of interest) and flow cytometry for mCherry or GFP. Furthermore, we consistently observed lower levels of GFP with the Ad-GFP-Zeb1 compared to Ad-GFP suggesting correlation between GFP and ZEB1 expression.

Flow cytometry

Fixed cells were permeabilized with phosphate buffered saline (PBS) containing 1% BSA and 0.1% Triton X-100 (TX100) and stained with primary antibodies in different combinations (mouse anti-MYH1, 1:300, MF20-c, DSHB; rat anti-mCherry, 1:500, M11217, Thermo Fisher; and rabbit anti-GFP, 1:500, ab290, Abcam) for 1 h in the dark on ice while shaking. After washing, cells were incubated with EdU Click-it reaction cocktail according to the manufacturer’s protocol (Invitrogen, C10419), and washed before incubation with secondary antibodies in different combinations (488-donkey anti-mouse, 1:200, A21202, Invitrogen; 555-donkey anti-rat, 1:200, Ab150154, Abcam; 555-donkey anti-mouse, 1:200, A31570, Invitrogen; and 488-donkey anti-rabbit, 1:200, A21206, Invitrogen) for 30 min in the dark on ice while shaking. Three final washes were performed in PBS/1% BSA/0.1% TX100 and Hoechst 33342 was added 5 min before flow cytometry using the LSRII flow cytometer (BD Biosciences). Data was analyzed using the FACSDiva software v8.0.1, and initially gated according to the CM marker MYH1 and then sub-fractionated based on the antibody amplified mCherry or GFP signal. Cells positive or negative for a reporter (mCherry or GFP) were gated according to EdU incorporation to determine cell cycle activity. Ploidy was addressed in subpopulations by Hoechst 33342 using gates (2n, 4, and > 4n) defined by the entire CM population. Each analysis as indicated consisted of at least three to nine independent experiments designated n, each comprising cells from one litter. Within an experiment, 1–3 replicates (n*) were performed and used as an average of the n for further statistical analysis as indicated.

RNA isolation, RNA integrity, and qRT-PCR

RNA was isolated from each sample and qRT-PCR was performed as described previously [3]. Briefly, the cells were lysed with TriReagent and the RNA was isolated using Polyacryl carrier, 1-Bromo-3-Chloro-Propane and 2-propanol. The RNA was rinsed using 75% ice cold ethanol. Finally, the RNA was dissolved in nuclease-free water and the RNA concentration was determined using a nano-drop. For qRT-PCR, cDNA was generated using the High-Capacity cDNA Reverse Transcriptase kit (Applied Biosystems; 4368814) according to the manufacturer’s recommendations. Each sample for qRT-PCR contained 2–4 ng cDNA (Supplementary Table 2) in a total volume of 10 µl and were analyzed in technical triplicates of qRT-PCR using a mixture of Power SYBRGreen PCR Master Mix (Applied Biosystems, 4367659) and appropriate forward and reverse primers (Supplementary Table 2). The qRT-PCR was run on a 7900HT Fast Real-time PCR system (Applied Biosystems) under the following conditions: holding for 10 min at 95 °C, hereafter 40 cycles consisting of 15 s of denaturation at 94 °C, 30 s of annealing at 57–60 °C (Supplementary Table 2) and 30 s of elongation at 72 °C. The obtained data was analyzed as previously described [3] by normalization to multiple stably expressed endogenous gene according to the qBase Plus 3.2 platform (Biogazelle).

Injection of adenovirus in P0 pups

The litter was gently taken from their home cage. Pups were then anesthetized by induction of hypothermia before 7.60 × 1014 PFU (in a total volume of 20 µl, diluted in sterile PBS) of the desired adenovirus was injected into the superficial temporal vein. Pups were reheated and placed together with their littermates before the litter was gently put back into their home cage. Following the pups received a 50 µl subcutaneous injections of EdU (2.5 mg/ml) at P4 and P6 before they were sacrificed by decapitation at P8. Hearts were dissected and either dissociated for flow cytometry as described above for P0 pups, except cells were strained (100 µm nylon cell strainer, 352360) prior to NBF fixation, and each heart was processed individually, or prepared for paraffin embedding (see below).

Immunohisto- and immunocyto-chemistry and microscopy

P8 hearts from virus injected pups were fixed in 10% NBF overnight, rinsed in PBS, dehydrated, and finally embedded in paraffin. Embedded specimens were then cut into 10 µm sections before mounting on glass slides and stored at 4 °C. Sections were deparaffinized and rehydrated before staining. Immunostainings were performed on NBF or PFA fixed cell cultures or paraffin embedded sections as previously described in Andersen et al. [2] with the following primary antibodies: rabbit anti-GFP (1:500, ab290, Abcam), mouse anti-MYH1 (1:300, MF20-c, DSHB), rabbit anti-ZEB1 (1:500, PA5-28,221, Thermo Fisher), rabbit anti-Mef-2c (1:500, 5030S, Cell Signaling Technology), mouse anti-actinin (1:200 A7811, Sigma), mouse anti-α-tubulin (1:250, 3873, Cell Signaling Technology), and 647-phalloidin (1:800, A30107, Thermo Fisher). Secondary antibodies used were: 488-donkey anti-rabbit (1:200, A21206, Invitrogen), 555-donkey anti-rabbit (1:200, A31570, Invitrogen), 647-donkey anti-mouse (1:200, A31571, Invitrogen). All sections were mounted with DAPI (Vectashield, Vector Lab., for Phalloidin and α-tubulin staining, Fluoroshield, Abcam was used). Microscopy was performed on a Leica DMI 4000 B microscope with a Leica CTR4000 illuminator and Leica DFC300FX/DFC 340 FX cameras, and confocal microscopy was performed on an Olympus FV1000MPE confocal laser scanning microscope equipped with an UPlanSApo 60x/1.20 water objective. During analysis, all camera settings and picture processing were applied equally to samples and controls.

Statistics and reproducibility

All statistics were performed using the GraphPad Prism (v 9.0.0) software and the appropriate tests, number of independent experiments (n) and replicates (n*) are defined in the corresponding figure legends. We used the significance level α = 0.05 for identifying significant results marked by asterisks, yet have indicated throughout the exact p-value if between 0.05 and 0.1 to enable objective evaluation of trends. For scRNA-seq, each timepoint includes three biologically independent experiments each comprising mouse pups derived from three distinct litters. Cells were pooled just before library generation and scRNA-seq. ScRNA-seq and subsequent gene expression analysis and transcription factor binding site analysis was performed using two different sequencing platforms (NextSeq and NovaSeq) with similar results. All transduction experiments were successfully reproduced with at least three biologically independent experiments obtaining similar results, thereby confirming the design and robustness.

留言 (0)

沒有登入
gif