Tuesday, February 13, 2018

AGBT: 10X Previews Three New Single Cell Applications

I spent breakfast with 10X Genomic's Michael Schnall-Levin and two of his 10X colleagues gave me a sneak peak at three new single cell products they are rolling out at the workshop I'm typing away at now.  These enable measuring protein targets of antibodies, mapping out accessible chromatin regions with ATAC-Seq, and mapping copy number variants (CNVs) at single cell resolution.  All use the existing Chromium Controller instrument.

In the first application, Single Cell Feature Barcoding, expression or T-cell receptor profiling can be coupled with readouts of DNA tags in the cells.  Those tags may be added to tag different mutants, such as DNA tags from large Cas9 screens, a technique called Perturb-Seq.  Another sort of feature of potential use are the use of inserted tags to record the lineage of cells.  Or whatever tagging logic scientists can design; now 10X can read these out simultaneously with 3' RNA-Seq or T-cell receptor sequences.  Alternatively, antibodies can be tagged with oligos and treated as features on the cells. 

One demonstrated example could provide a broad new window into T-cell function.  It is possible to construct complexes which will engage T-cell receptors in place of proper antigen presenting cells. By using the 10X workflow and antibodies which report the antigens in these complexes, it is possible to simultaneously read out the binders of T-cell receptors and the full length sequences of both chains of the T-cell receptor.

ATAC-Seq uses a hyperactive transposon which preferentially inserts in open chromatin which in turn is highly correlated with active transcription start sites.  10X's solution provides a parallel method to single cell RNA-Seq.  In some sense this is like the above features, though the feature is added to the cell after encapsulation.  With 10X's method, populations of 10,000 to 100,000 cells can be profiled by ATAC-Seq. 

Copy number variation has attracted great attention, particularly in the context of cancer.  Tumors often have unstable chromosomes, leading to deletions and duplications.  Selective pressures can enrich for duplications containing oncogenes or for deletions spanning tumor suppressors.  A wide array of technologies have been developed for CNV determination, but these are nearly always used on bulk samples.  Even most cancer cell lines are known to be heterogeneous for copy numbers.

The CNV method uses a new protocol for single cell encapsulation on the 10X platform.  In the older version, cells were captured in droplets with the gel beads bearing molecular barcodes.  Now, a first pass through the instrument captures cells in their own little solid beads.  These "cell beads" can be processed to lyse the cells, remove cellular proteins and denature the DNA with alkali, but the DNA remains trapped in the bead.  A second pass through the instrument forces one-to-one pairings of a cell bead and a gel bead into droplets for further processing.  10X calls this new process CGGBs -- cell beads, gel beads.This opens up the possibility of applying treatments to the encapsulated cells that may not be compatible with the barcode-bearing beads.

Compared to lower throughput, plate-based methods such as DOP-PCR or MALBEC, 10X's CNV workflow shows more uniform coverage of the genome, enabling more sensitive detection of copy number variants.   In a spike-in experiment with 5% of the second cell line, 10X could successfully detect the rare copy number variant.  

Applying this technique to a cell line "BJ' thought to be homogeneous revealed a consistent CNV present at 7% frequency.  In the widely used COLO-869 cell line, this technique found great heterogeneity.  For a COLO-869 Chromosome 1 CNV that has been reported previously in COLO-869, four distinct subclones differed in the precise breakpoints.  On a breast cancer sample, the method can distinguish normal cells from cancer cells and detect and quantify focal amplifications of specific driver genes. 

All of these new products will be supported by analysis tools within 10X's software environment.  


No comments: