Journal: Nature Communications

Article Title: Integrative genotyping of cancer and immune phenotypes by long-read sequencing

doi: 10.1038/s41467-023-44137-7

Figure Lengend Snippet: a Intermediate single-cell cDNA libraries of cancer specimens are used for targeted amplification of transcripts carrying natural barcodes. By sequencing long amplicons on the Oxford Nanopore platform multiple genetic barcodes such as somatic nuclear mutations can be detected from the same amplification product. Applications for long-read sequencing include the read-out of T cell receptor (TCR) or CAR sequences, somatic nuclear and mitochondrial DNA mutations, fusion transcripts, and alternative splicing events (gene isoforms). The genotyping information from long-read sequencing is integrated with the gene expression data from short-read sequencing. b Amplicons are generated using a 3-step PCR. In the first PCR template-switch oligo (TSO) artifacts are removed with generic amplification of cDNA using a biotinylated 3′ primer and streptavidin purification. In the second PCR, gene-specific biotinylated 3′ primers are used to amplify loci of interest. After a second streptavidin purification, target genes are amplified with nested gene-specific 3′ primers to provide sufficient material for sequencing. c Overview of the nanoranger workflow. Multimer reads are deconcatenated by identifying transcripts with alignment against a reference transcriptome (1). After extraction of subreads (2), cell barcodes are identified (3) and TCR information is processed or transcripts are genome-aligned (4) for downstream genotyping. d Examples of gene coverage with (black) and without (gray) removal of TSO artifacts. The blue line indicates the primer binding site and the red ribbon shows locations of mutations used for lineage tracking in AML detected with each primer set. e Benchmarking of multimer demultiplexing using nanoranger and longbow on artificially generated multimers using the ISO-MAS-seq protocol. Of note, nanoranger deconcatenates multimer reads agnostic of adapters between transcripts, while longbow is optimized for known adapter sequences.

Article Snippet: Together, ONT-based long-read sequencing of targets amplified from intermediate 10x Genomics cDNA libraries appears to be best suited for targets on highly expressed, short transcripts within the first 4000 bp of the 5′ end.

Techniques: Amplification, Sequencing, Expressing, Generated, Purification, Extraction, Binding Assay

Journal: Nature Communications

Article Title: Integrative genotyping of cancer and immune phenotypes by long-read sequencing

doi: 10.1038/s41467-023-44137-7

Figure Lengend Snippet: a Tumor-infiltrating T cells from one melanoma case (patient C described in Oliveira et al., Nature 2021 ) were isolated and used for single cell sequencing. After generation of enriched T cell receptor (TCR) libraries, they were either fragmented and sequenced with Illumina (left) or directly sequenced without fragmentation with Oxford Nanopore (ONT) (right). b Comparison of reads obtained with Illumina and ONT ( nanoranger ) per cell barcode shown for TCRα (light) and TCRβ (dark) sequences (left) or number of cells obtained with Illumina and ONT with a particular CDR3α (light) or CDR3β (dark) (right). The rectangle indicates unproductive CDR3s filtered by cellranger. c UMAP representation of cell types (top left), cells with detectable TCR (top right), expression of CD14 and CD3E (bottom). The bar plot demonstrates the number of TCR sequences obtained by Illumina and ONT sequencing. d Overview mixing experiment with Kasumi-1 (acute myeloid leukemia, AML) and K562 (chronic myeloid leukemia, CML) cells. Kasumi-1 contain a homozygous TP53 R248G mutation and the RUNX1::RUNX1T1 fusion gene. K562 express only one TP53 allele with a truncating frameshift mutation ( TP53 Q136fs ) and BCR::ABL1 . Both TP53 mutations are detected with the same primer. e Absolute number and percentage of genotyped Kasumi-1 (red) and K562 cells (yellow) for the TP53 R248G mutation (left) and percentage of cells carrying either TP53 R248G or TP53 Q136fs (right). f Expression of TP53 and percentage of cells with detectable TP53 transcripts (left). Number of cells genotyped as function of the number of TP53 reads shown for Kasumi-1 (red) and K562 (yellow) in a downsampling experiment (right). g Percentage of genotyped cells for BCR::ABL1 and RUNX1::RUNX1T1 (left). Percentage of cells with BCR::ABL1 and RUNX1::RUNX1T1 shown for Kasumi-1 (red) and K562 (yellow) (right). h Coverage across all detectable genes from a 10x Genomics cDNA library after removal of TSO artifacts as function of transcript length.

Article Snippet: Together, ONT-based long-read sequencing of targets amplified from intermediate 10x Genomics cDNA libraries appears to be best suited for targets on highly expressed, short transcripts within the first 4000 bp of the 5′ end.

Techniques: Isolation, Sequencing, Comparison, Expressing, Mutagenesis, cDNA Library Assay

Journal: Nature Communications

Article Title: Integrative genotyping of cancer and immune phenotypes by long-read sequencing

doi: 10.1038/s41467-023-44137-7

Figure Lengend Snippet: a Experimental workflow of comparison between nanoranger and genotyping of transcriptomes (GoT). A pretreatment bone marrow sample of AML1022 at relapse after allogeneic hematopoietic stem cell transplantation (HSCT) was used for single cell cDNA library preparation according to the standard 10x Genomics 5′ gene expression protocol and following the modified 5′ GoT protocol with in-droplet inclusion of gene-specific reverse transcriptase primers. Both cDNAs were taken forward for sequencing with the standard nanoranger protocol (orange), GoT using Illumina sequencing (black) and GoT using Oxford Nanopore sequencing (blue). b , c Number of cells genotyped with each experimental condition (b) and percentage of genotyped cells across hematopoietic differentiation states ( c ). d Comparison of apparent single cell variant allele frequencies (VAFs) for SF3B1 K700E in donor- versus recipient-derived cells to demonstrate specificity of genotyping with each experimental condition. e Comparison of cell barcodes identified with each condition. The venn diagrams demonstrate the number of cell barcodes that are uniquely identified or shared across experimental conditions. To enable direct comparison of captured cell barcodes, the cDNA for the GoT condition was used as input for nanoranger . f Minimal read length versus number of reads for cell barcodes identified with GoT on Illumina and ONT (black) versus those identified only with GoT on ONT (blue), demonstrating the preferential sequencing of shorter fragments with Illumina sequencing.

Article Snippet: Together, ONT-based long-read sequencing of targets amplified from intermediate 10x Genomics cDNA libraries appears to be best suited for targets on highly expressed, short transcripts within the first 4000 bp of the 5′ end.

Techniques: Comparison, Transplantation Assay, cDNA Library Assay, Expressing, Modification, Sequencing, Nanopore Sequencing, Variant Assay, Derivative Assay

Journal: Nature Communications

Article Title: Integrative genotyping of cancer and immune phenotypes by long-read sequencing

doi: 10.1038/s41467-023-44137-7

Figure Lengend Snippet: a Co-existence of two subclones in de-novo AML 1. Detection of the somatic nuclear mutations NPM1 W287fs (clone 1) and NPM1 W288fs (clone 2) demonstrates co-existence of two AML clones that differ in the presence of FLT3-ITD and loss of heterozygosity on chromosome 13 ( loh(13) ) as well as several mitochondrial DNA mutations. b Differential analysis of bulk mitochondrial DNA heteroplasmy between clone 1 and clone 2 in de-novo AML 1. c Heatmap demonstrating molecular features of clone 1 and clone 2 in 525 cells of de-novo AML 1. d , e Differential gene expression analysis (DGEA) between GMP-like cells of clone 1 and clone 2. f Comparison of BCR::ABL1 amplicons in two Philadelphia + (Ph + ) acute lymphoblastic leukemia (ALL) cases, in K562 cells and Ph + chronic myeloid leukemia (CML). In all 4 cases the same ABL1 -specific primer was used. The p190 variant in ALL produces a shorter fusion transcript than the p210 variant in CML making it more detectable with targeted long-read sequencing from 10x Genomics cDNA libraries. g Identification of BCR::ABL1 + cells in ALL bone marrow. UMAP plots show cell type annotation (left), samples (middle) and detection of BCR::ABL1 transcripts (right) in ALL1 (black) and ALL2 (red). h BCR::ABL1 + cells in bone marrow of ALL1 and ALL2 (left) and cells with detectable CNV changes in re-analyzed ALL datasets from refs. and (right), mapped to a healthy bone marrow reference.

Article Snippet: Together, ONT-based long-read sequencing of targets amplified from intermediate 10x Genomics cDNA libraries appears to be best suited for targets on highly expressed, short transcripts within the first 4000 bp of the 5′ end.

Techniques: Clone Assay, Expressing, Comparison, Variant Assay, Sequencing

Journal: Nature Communications

Article Title: Integrative genotyping of cancer and immune phenotypes by long-read sequencing

doi: 10.1038/s41467-023-44137-7

Figure Lengend Snippet: a Targeted amplification of PTPRC to detect differential splicing of exon 4 which determines expression of CD45RA (exon 4 expressed) versus CD45RO (exon 4 not expressed) (top). Targeted amplification dramatically increases coverage of PTPRC (red) compared to whole-transcriptome (WT) amplified cDNA (gray), both sequenced on the Oxford Nanopore platform. b UMAP representation of tumor-infiltrating T cells (TILs) and circulating T cells from melanoma Patient C (Oliveira et al., Nature 2021 ) . The top row shows expression of PTPRC exon 4 (CD45RA) and the bottom row shows CD45RA protein expression measured by CITE-seq. c Expression of PTPRC (exon 4) (left) and CD45RA measured by CITE-seq (right) across T cell subsets in bone marrow of AML1007 before infusion of ipilimumab (baseline) and after 1 or 4 cycles of ipilimumab. d Targeted amplification of CTLA-4 to detect exon 3 which discriminates the soluble (exon 3 absent) and membranous (exon 3 present) isoforms (top). Knee plot demonstrating the high degree of enrichment of CTLA-4 transcripts with targeted long-read sequencing (red) versus whole-transcriptome long-read sequencing (gray) (bottom). e Expression of soluble (black) and membranous (yellow) isoforms of CTLA-4 across AML3005, tumor-infiltrating T cells of Patient C, and a CAR T cell infusion product (top). Expression level of CTLA-4 as measured by short-read sequencing indicates specific detection with the targeted approach (bottom). f Distribution of percentage of reads with membranous CTLA-4 across eight different T cell single-cell cDNA libraries for a total of 4786 cells. Patients C and D were previously described by ref. . CAR IP - CAR T cell infusion product.

Article Snippet: Together, ONT-based long-read sequencing of targets amplified from intermediate 10x Genomics cDNA libraries appears to be best suited for targets on highly expressed, short transcripts within the first 4000 bp of the 5′ end.

Techniques: Amplification, Expressing, Sequencing