FREQUENTLY ASKED QUESTIONS

If you have questions, comments or suggestions that are not addressed, here please contact us.

CITE-seq count FAQ. Can be found in the GitHub wiki.

How do I pronounce ECCITE-seq?  Like you’re EXCITED to be doing it

How do I pronounce ASAP-seq?  Ay-Sap-seq.  We take no responsibility if this encourages people to mispronounce ATAC-seq (Attack = correct, Ay-Tac = incorrect).  Of course, “-seq” is always pronounced “seek”!

Should I use Streptavidin-Biotin conjugation or direct covalent conjugation of antibodies to oligonucleotides? Antibody-oligo conjugates can be prepared by various methods which form covalent links between antibodies and oligos (e.g. amine-reactive NHS or iEDDA-, Maleimide- click chemistries) or by indirect streptavidin-biotin-linkage. For our proof- of-principle experiments we used streptavidin-biotin-linkage to couple oligos to antibodies, and include a cleavable linker in the oligos. We are now using direct antibody-oligo conjugation chemistry (iEDDA, essentially as described in Van Buggenum et al., Scientific Reports, 2016; without a cleavable linker), and antibody-oligo conjugates from BioLegend. We have compared direct and indirect conjugation methods and had very comparable results. Direct conjugation has the added benefit that larger panels can be pooled and stored for prolonged periods of time.

How many antibodies can be multiplexed in one CITE-seq experiment? The DNA-barcode on the antibody-oligos allows a virtually ‘limitless’ number of barcodes, far exceeding the numbers of fluorophores for flow cytometry or isotopes for mass cytometry. We have successfully used ~125 antibodies and do not foresee any reason why this number can’t be further increased. We have also not observed any competition between mRNAs and Antibody-oligos for polyT primers, larger panels and high concentrations of Cell Hashing antibodies did not result in decreased nUMI or nGene counts.

Do antibodies have to be pooled always shortly before the run? Can’t I store them in a pooled panel? Yes you can make and store CITE-seq panels! Once we know the best titrated concentration for each antibody we typically pool all antibodies at appropriate concentrations into larger panels, given these are directly and irreversibly covalently linked antibody-oligo conjugates (e.g. TotalSeq reagents). When dealing with Streptavidin-Biotin conjugates we typically only pool them shortly before the run and never store these as pooled panels.

Why did you chose a small RNA read 2 PCR handle on the CITE-seq antibody-oligos?  We wanted to keep the oligos on the antibodies as short as possible to avoid potential adverse effects on antibody specificity or accessibility. Small RNA read 2 handle is currently the shortest sequencing library handle from Illumina without partial reverse complementarity to read 1, as for Truseq DNA or Nextera primers. Illumina sequencers run with a pool of all possible illumina sequencing primers so the machines are agnostic to the primers used in a particular library pool. Therefore ADT/HTO and cDNA libraries can be pooled without any alterations to the standard sequencing protocol.

What do the “xxxx” mean in the RPI-x and TruSeq D7xx_s primers?  The “x” nucleotides are meant to indicate that you can use whatever sequence you want in these positions.  Standard index sequences from Illumina work, as do custom sequences.  The 10x i7 indexes used for indexing your 3′ tag cDNA libraries are a pool of 4 separate indexes per sample, so be sure to check that the indexes you use for your ADT / HTO libraries have sufficient edit distance from these indexes and from each other.

What is with the “_s” suffix on the TruSeq D7xx_s primers?  The “_s” suffix indicates that these primers are “short”.  Their 3 end is right at the start of the 12 nt stretch that is common to TruSeq read 1 and TruSeq read 2 primers.  The “_s” primers were designed this way to prevent mis-priming on the read 1 end of the HTO library. Libraries made with these _s primers are not truncated in any way, they will still sequence normally on all machines.  Refer to the Cell_Hashing_assay_scheme for more information.  We have seen better performance with these “D70X_s” vs. “D70X” primers but also never went back to really quantify it.

I want to pool different ADT and/or HTO libraries but protocol only lists one D701 and RPI1 primer, which other primers can I use? We often pool multiple experiments in one sequencing run. For this, ADT and/or HTOs can be amplified with different RPI and/or D7 primer sets from Illumina.  Please refer to the Illumina Adapter Sequences Document.  The index sequence (bold sequence in the protocol) can be exchanged for any other sequence that is compatible with your assay.

My CITE-seq protein tag libraries have a 6 nt index but my cDNA libraries have an 8 nt index.  What should I do?  Illumina’s RPI-x primers are pretty old and harken back to the days when 6 nt was plenty to index the numbers of samples that people would multiplex.  We design our own RPI-x primers with 8 nt indexes, but if you have used the Illumina sequences with 6 nt indexes and want to pool with libraries with 8 nt i7 indexes, don’t worry!  The next two nucleotides after the index position are “AT” so if you simply add an “AT” after the index sequence on the sample sheet it will demux appropriately.  For example, on the sample sheet, the RPI-1 index: “ATCACG” becomes “ATCACGAT”.

I want to order my own RPI-x primers.  Which bases can I change?  The RPI-x primer contains the P7 sequence, then the index (xxxxxx) followed by the read 2 primer binding site.  You can place your desired indices in place of the “x” nucleotides on the sequence below.  The original Illumina primers have 6 nt indices, we typically use 8 nt.

RPI-x:  CAAGCAGAAGACGGCATACGAGATxxxxxxGTGACTGGAGTTCCTTGGCACCCGAGAATTCCA  

Where should I order the primers described in the protocol and what should they be diluted in? We order our oligonucleotides at any common oligonucleotide synthesis company. We typically dilute our oligo stocks to 100 µM in water or TE, and have our working concentrations as described in the protocols.

Do any of the oligos for the protocol have to be ordered PAGE or HPLC purified?  We typically do not order our oligos HPLC or PAGE purified.  Desalted oligos work well in our hands.

Why are you adding ADT and/or HTO additive primers to the cDNA amplification?  Addition of an antibody-oligo specific primer (cDNA additive primer) during cDNA amplification at low concentration significantly improves ADT and/or HTO library purity and yield in the subsequent ADT-specific library PCR. We add these primers in very low concentration with the goal of getting some amplification, but not enough to interfere with the amplification of the full length mRNA-derived cDNAs.

CD45 vs. CD298 & B2M Hashing Antibodies. For our proof of principle Cell Hashing experiments using PBMCs we have used multiple ‘ubiquitous’ immune markers including CD45. For human cells we have now entirely switched to using a mix of CD298 and B2M, which are expressed on a large variety of tissues and cell types, including immune cells. This allows Cell Hashing of virtually any sample, which might contain mixtures different cell types.  We note that a recent paper described exactly this combination of markers as a universal live cell barcoding reagent for CyTOF.  The CD298 / B2M combination is also what is sold by BioLegend for hashing of human samples.

My ADT (or HTO) library contains a large fraction of RT-TSO dimer (~150nt). The RT-TSO dimer is carried over from the cDNA amplification and is typically dwarfed simply by doing 1 or 2 additional cycles of ADT PCR on the purified ADT library. Carryover of cDNA amplification primers will re-amplify the RT-TSO dimer and mRNA-cDNA fragments in the ADT library, if this is a recurrent problem try purifying your ADT library with two rounds of 1.8X SPRI (instead of 2X SPRI). Also, this is typically a bigger problem with very small antibody panels (less than ten), the larger your antibody panel gets the less prominent this dimer will be.

My ADT (or HTO) library contains a large broad peak at around 400 bp. Excessive library amplification leads to exhaustion of primers (and/or dNTPs) which results in “daisy-chains” or “bubble products”, heterogenous library products that anneal by their P5 and/or P7 adapter sequences. These appear as very characteristic high molecular weight smooth DNA peaks typically around double the size of the correct product and can be more or less prominent (panel A,B below). This larger peak will cluster and sequence perfectly and it will be also a good quality library. But it is very hard to accurately quantify these libraries. If you encounter a bubble product in your library you can get rid of it by simply performing another PCR reaction (using same primersets as first PCR or P5/P7 generic primers) for one or two cycles on the 1.6X SPRI cleaned up product. Depending on the amount of ‘daisy-chained’ product, these additional PCR cycles will convert the library partially (panel C) or entirely (panel D) to double stranded DNA, collapsing the product size to the correct size range.

I do not observe a ADT (or HTO) product on the Bioanalyzer in the supernatant fraction after cDNA amplification. We typically never run the ADT/HTO-supernatant fraction on the BA, there is usually nothing visible from the ADT (or HTOs)! This fraction typically only contains a very prominent ~100 band from RT-TSO hybrid (see assay our assay scheme). After the ADT/HTO PCR there should be a clear and sharp product at ~180 bp.

What are P5 and P7 generic primers? These primers anneal to the P5 and P7 regions of Illumina libraries and can be used to re-amplify any Illumina sequencing library.  P5_generic:  5’AATGATACGGCGACCACCGAGATCTACAC, P7_generic: 5’CAAGCAGAAGACGGCATACGAGAT

Are 5P CITE-seq and 5P Hashing antibodies commercially available? Yes.  5P compatible antibodies for CITE-seq and Hashing are commercially available as TotalSeq-C reagents, Please note that these have different amplification handles than the sequences we described in the ECCITE-seq paper.  Details of how to use these reagents and recover tags can be found here.  

I want to run FACS prior to CITE-seq. Can I stain samples with Cell Hashing and/or CITE-seq and our fluorescent panel at the same time? Yes. We often first enrich cell populations by FACS and then run CITE-seq or multiplex different sorted samples by Cell Hashing. To maximize viability and eliminate multiple staining/incubation steps we typically stain the cells with CITE-seq and/or Cell Hashing and flow cytometry antibodies before the sort and then go directly into CITE-seq after the sort. Although the sorting process also functions as a wash to some extent we still wash the cells twice before going into the scRNA-seq run.

I am worried about including 0.01% Tween in the staining buffer. Is it essential? Tween is optional. We use a very low concentration that in our hands does not interfere with cell viability but allows cells to pellet better during washing steps. We found that 0.01% Tween is particularly useful when dealing with low cell numbers. We have not seen effects on viability on PBMCs or different cell lines. We recommend testing different washing buffer and centrifugation speed conditions when in doubt.

Do TotalSeq A reagents work on 10x Genomics 3P v3 kits? YES!  Although 10x doesn’t support them (for their own reasons), we have tested our polyA antibodies on 10x Genomics 3P version 3 and they worked as well as on version 2. The mRNA capture oligos on 10x Genomics 3P v2 and v3 are functionally almost identical. They have the same amplification handle, and most importantly both have a polyT capture sequence. There are small differences in UMI length (12 vs 10 nt) and anchor. Therefore home-made polyA conjugates and the well benchmarked TotalSeq-A reagents work equally well on 10x Genomics v3 and have the added benefit that they are be compatible with other current (and future) polyA-based single cell RNA-sequencing platforms.

My ADT and/or hashtag reads have low quality scores.  Is this a problem?  NO.  Low AVERAGE Q scores for read 2 on hashtag and ADT libraries are expected.  For polyA CITE-seq / hashing reagents and TotalSeqA, once you get past the barcode sequence, you go straight into a stretch of A nucleotides, which causes the quality to drop.  Also, if you are using a long ~150 cycle read for read 2, you’ll sequence to the end of the molecule and the quality will drop even more as there are no more bases to read.  The actual quality of the bases that matter (the barcode) is usually great.

Can I use TotalSeq-C reagents for ASAP-seq?  Maybe.  It works to some degree, but it isn’t optimal for a couple of reasons.  Firstly, the 13 nt annealing stretch on the TotalSeq-C reagents has a much lower annealing temperature than either the polyA stretch on TotalSeq-A or the 10x capture sequence 1 on TotalSeq-B.  This makes the bridging reaction less efficient, and while we can get it to work, it is not as effective as using A or B conjugates.  Secondly, the amplification handle on TotalSeq-C is the Nextera read 2 handle, so the protein tags from bridged TotalSeq-C reagents have the same handles as the transposase-generated ATAC fragments, making it not possible to separately amplify them.  All of this notwithstanding, if you have a large library of TotalSeq-C reagents and you are committed to trying this, we’re happy to share our experiences if you want to contact us.  

What is the UBI in ASAP-seq and how does it work?  Unlike scRNA-seq, scATAC-seq methods typically do not feature Unique Molecular Identifiers (UMIs).  This is because the genomic coordinates of fragments are usually unique so UMIs are unnecessary.  This poses a problem for capturing protein tags that themselves do not have UMIs (E.g. TotalSeq-A) as it becomes impossible to distinguish PCR duplicates from separately captured tags.  In ASAP-seq, a Bridge oligo acts as a template to convert the antibody-bound oligo into a sequence that can anneal to the ATAC-seq barcoding oligo.  By adding a random stretch of nucleotides into the bridge oligo, each individual bridging event creates a unique sequence that can be used for tag counting.  Since each antibody:oligo can only be extended once, the UBI serves as a substitute for a UMI and allows accurate quantification of protein tags.