Limits...
T cell fate and clonality inference from single cell transcriptomes

View Article: PubMed Central - PubMed

ABSTRACT

The enormous sequence diversity within T cell receptor (TCR) repertoires allows specific TCR sequences to be used as lineage markers for T cells that derive from a common progenitor. We have developed a computational method, called TraCeR, to reconstruct full-length, paired TCR sequences from T lymphocyte single-cell RNA-seq by combining existing assembly and alignment programs with “combinatorial recombinome” sequences comprising all possible TCR combinations. We validate this method to quantify its accuracy and sensitivity. Inferred TCR sequences reveal clonal relationships between T cells whilst the cells’ complete transcriptional landscapes can be quantified from the remaining RNA-seq data. This provides a powerful tool to link T cell specificity with functional response and we demonstrate this by determining the distribution of members of expanded T cell clonotypes in a mouse Salmonella infection model. Members of the same clonotype span early activated CD4+ T cells, as well as mature effector and memory cells.

No MeSH data available.


Distributions of lengths of reconstructed TCR sequences. Reconstructed sequences were trimmed to include the region derived from the V gene, junction and J gene. The lengths of these sequences are plotted as histograms and kernel density estimates for TCRa (upper) and TCRb (lower). Dotted lines represent the interquartile range of lengths of full-length sequences derived from the combinatorial recombinome files.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC4835021&req=5

Figure 1: Distributions of lengths of reconstructed TCR sequences. Reconstructed sequences were trimmed to include the region derived from the V gene, junction and J gene. The lengths of these sequences are plotted as histograms and kernel density estimates for TCRa (upper) and TCRb (lower). Dotted lines represent the interquartile range of lengths of full-length sequences derived from the combinatorial recombinome files.

Mentions: Our method (Supplementary Fig.1a) extracts TCR-derived sequencing reads for each cell by alignmentagainst ‘combinatorial recombinomes’ comprising all possiblecombinations of V and J segments (Supplementary Fig. 1b). Reads are then assembled into contiguoussequences which are analysed to find those that represent full-length, recombinedTCR sequences. Importantly, the reconstructed recombinant sequences typicallycontain nearly the complete length of the TCR V(D)J region (Fig. 1) and so allow high-confidence discrimination betweenclosely related and highly-similar gene segments. Here, we use scRNA-seq datagenerated using the SMART-Seq protocol26 with the Fluidigm C1 microfluidics system. Our method would,however, work with any scRNA-seq data derived from full-length cDNA.


T cell fate and clonality inference from single cell transcriptomes
Distributions of lengths of reconstructed TCR sequences. Reconstructed sequences were trimmed to include the region derived from the V gene, junction and J gene. The lengths of these sequences are plotted as histograms and kernel density estimates for TCRa (upper) and TCRb (lower). Dotted lines represent the interquartile range of lengths of full-length sequences derived from the combinatorial recombinome files.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC4835021&req=5

Figure 1: Distributions of lengths of reconstructed TCR sequences. Reconstructed sequences were trimmed to include the region derived from the V gene, junction and J gene. The lengths of these sequences are plotted as histograms and kernel density estimates for TCRa (upper) and TCRb (lower). Dotted lines represent the interquartile range of lengths of full-length sequences derived from the combinatorial recombinome files.
Mentions: Our method (Supplementary Fig.1a) extracts TCR-derived sequencing reads for each cell by alignmentagainst ‘combinatorial recombinomes’ comprising all possiblecombinations of V and J segments (Supplementary Fig. 1b). Reads are then assembled into contiguoussequences which are analysed to find those that represent full-length, recombinedTCR sequences. Importantly, the reconstructed recombinant sequences typicallycontain nearly the complete length of the TCR V(D)J region (Fig. 1) and so allow high-confidence discrimination betweenclosely related and highly-similar gene segments. Here, we use scRNA-seq datagenerated using the SMART-Seq protocol26 with the Fluidigm C1 microfluidics system. Our method would,however, work with any scRNA-seq data derived from full-length cDNA.

View Article: PubMed Central - PubMed

ABSTRACT

The enormous sequence diversity within T cell receptor (TCR) repertoires allows specific TCR sequences to be used as lineage markers for T cells that derive from a common progenitor. We have developed a computational method, called TraCeR, to reconstruct full-length, paired TCR sequences from T lymphocyte single-cell RNA-seq by combining existing assembly and alignment programs with “combinatorial recombinome” sequences comprising all possible TCR combinations. We validate this method to quantify its accuracy and sensitivity. Inferred TCR sequences reveal clonal relationships between T cells whilst the cells’ complete transcriptional landscapes can be quantified from the remaining RNA-seq data. This provides a powerful tool to link T cell specificity with functional response and we demonstrate this by determining the distribution of members of expanded T cell clonotypes in a mouse Salmonella infection model. Members of the same clonotype span early activated CD4+ T cells, as well as mature effector and memory cells.

No MeSH data available.