The below diagram illustrates the flow of a set of cell suspension samples through the process of library preparation and sequencing. A cell suspension, which can either be its own unique sample or (optionally) a pool of multiplexed samples, is used to generate a single cDNA library. A set of individually barcoded libraries will then be pooled together for sequencing on an Illumina flow cell. The differences between these two types of pooling (pooling samples at the start for a multiplexed library and pooling libraries at the end for multiplexed sequencing) are discussed in detail below.
S.V. Medaris/Media Solutions
In a 10X Genomics single cell RNA-seq library preparation, there are two points where samples may be pooled together. The first point (1) is pooling of samples for the initial input into the 10X Genomics instrument for library preparation – pooling for sample multiplexing. There are currently three main ways of accomplishing this:
- “Hashing” or “hashtagging” of samples. In this approach, the cells of each individual starting sample are stained with distinct antibody-nucleotide barcode conjugates (such as the TotalSeq antibodies available from BioLegend). Since we know that Sample #1 was stained with antibody #1 and Sample #2 was stained with antibody #2, and we know the sequences of the unique barcodes linked to each antibody, we can now pool together Sample #1 and Sample #2 into a single sample (“Sample A”). In the sequencing data, we will be able to see which cells have the antibody #1 barcode and which cells have the antibody #2 barcode, so that we can assign them to Sample #1 or Sample #2.
- Built-in probe barcodes in the fixed cell “Flex” assay. Rather than using polyA-based capture of mRNA as in the fresh cell assays, this assay relies on the tagging of transcripts with oligonucleotide probes designed to be complementary to the sequences of approximately 19,000 transcripts in either human or mouse. The probes also contain a separate barcode sequence; in some versions of the assay, the reagent kits will include multiple versions of the probe set reagent that each have a unique sequence for this barcode (e.g. barcodes 1 through 4). As in the “hashing” approach, by using the probe set with barcode #1 to label Sample #1 and using the probe set with barcode #2 to label Sample #2, these samples can now be combined into a single sample for library preparation.
- “On-chip multiplexing” (OCM), a new and distinct version of 10X’s main 3′ and 5′ Gene Expression assays. The specifics of this approach are discussed separately here for brevity; in short, this method allows you to combine groups of up to four samples into a single pooled sample for library preparation by loading them in a specific order onto the microfluidic chip used in the assay, without needing to do any separate labeling of the sample with a unique barcode.
Though approaches #2 and #3 are intentionally designed to essentially require pooling of samples for sample multiplexing, this type of pooling is always optional. The majority of 10X Genomics projects in the GEC do not use any form of sample multiplexing for the library construction input. While sample multiplexing can significantly cut reagent costs for the project, each of those three options has drawbacks or limitations that make them unsuited for certain experiments.
The second point (2) where samples can be pooled in a 10X Genomics experiment is at the end of the process – pooling of libraries for sequencing. This pooling step is always done in virtually every type of RNA-seq library in order to facilitate parallel sequencing of large numbers of samples. This is possible because in every type of 10X Genomics library preparation, the final step is a “Sample Index PCR” reaction where the library is amplified by PCR using universal primers that also add a unique pair of index sequences separately to every library – e.g. in the diagram above, Libraries A through D will all receive a different known index sequence added to every molecule in the library. If this step was not included, these libraries would have to be sequenced separately in their own individual lane of the sequencing flow cell shown at the bottom of the diagram. By pooling them together like this, the uniquely indexed libraries can be sequenced together and distributed across every lane of the flow cell, allowing us to control for small variations in sequencing performance from one lane to the next.
To summarize:
(1) Pooling of samples for sample multiplexing is an optional step. The primary benefit to this is reducing the cost of the experiment by reducing the required number of units of expensive library preparation reagents.
(2) Pooling of libraries for sequencing is always done because it is expressly built into every library preparation protocol. This allows sequencing to be cheaper and more uniform across different samples. When a GEC staff member provides you with a project quote that includes sequencing, that sequencing estimate will cover the entire project (i.e. it is not a per sample cost) because we know that the libraries will be pooled into a single object to be loaded onto the sequencer.