Depletion of Abundant Sequences by Hybridization (DASH): using Cas9 to remove unwanted high-abundance species in sequencing libraries and molecular counting applications.
Gu W, Crawford ED, O'Donovan BD, Wilson MR, Chow ED, Retallack H, DeRisi JL. · 2016
Read Depletion of Abundant Sequences by Hybridization (DASH): using Cas9 to remove unwanted high-abundance species in sequencing libraries and molecular counting applications. by Gu W, Crawford ED, O'Donovan BD, Wilson MR, Chow ED, Retallack H, DeRisi JL. (2016) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc4778327.txt.
Introduction
W. Gu and six co-authors introduce a clean-up trick for DNA sequencing in Genome Biology in 2016. Sequencing runs waste much of their capacity on a few over-common RNA species, drowning the rare signals researchers actually want. Their method, DASH, enlists the Cas9 gene-cutting protein, loaded with guide RNAs that target the unwanted species for cutting before the run.
Cut species cannot take up sequencing space, so rare ones stand out. In human cells the method cut a dominant mitochondrial RNA by more than 99 percent. In patient samples it enriched germ sequences, easing diagnosis, and the team shows a cancer use as well. Because the guides can be rewritten for any target list, the trick adapts to many sample types at no added run cost.
Catalog record
| CREATOR | Gu W, Crawford ED, O'Donovan BD, Wilson MR, Chow ED, Retallack H, DeRisi JL. |
|---|---|
| YEAR | 2016 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 58 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Depletion of Abundant Sequences by Hybridization (DASH): using Cas9 to remove unwanted high-abundance species in sequencing libraries and molecular counting applications By W Gu, E D Crawford, B D O’Donovan, M R Wilson, E D Chow, H Retallack, J L DeRisi Published in: Genome Biology DOI: 10.1186/s13059-016-0904-5 ABSTRACT Abstract Next-generation sequencing has generated a need for a broadly applicable method to remove unwanted high-abundance species prior to sequencing. We introduce DASH (Depletion of Abundant Sequences by Hybridization). Sequencing libraries are ‘DASHed’ with recombinant Cas9 protein complexed with a library of guide RNAs targeting unwanted species for cleavage, thus preventing them from consuming sequencing space. We demonstrate a more than 99 % reduction of mitochondrial rRNA in HeLa cells, and enrichment of pathogen sequences in patient samples. We also demonstrate an application of DASH in cancer. This simple method can be adapted for any sample type and increases sequencing yield without additional cost. Electronic supplementary material The online version of this article (doi:10.1186/s13059-016-0904-5) contains supplementary material, which is available to authorized users. Keywords: Cas9, CRISPR, Depletion, Sequencing, RNA-Seq, Infectious Disease, Cancer, Diagnostics BACKGROUND The challenge of extracting faint signals from abundant noise in molecular diagnostics is a recurring theme across a broad range of applications. In the case of RNA sequencing (RNA-Seq) experiments specifically, there may be several orders of magnitude difference between the most abundant species and the least. This is especially true for metagenomic analyses of clinical samples like cerebrospinal fluid (CSF), whose source material is inherently limited [1], making enrichment or depletion strategies impractical or impossible to employ prior to library construction. The presence of unwanted high-abundance species, such as transcripts for the 12S and 16S mitochondrial ribosomal RNAs (rRNAs), effectively increases the cost and decreases the sensitivity of counting-based methodologies. The same issue affects other molecular clinical diagnostics. In cancer profiling, the fraction of the mutant tumor-derived species may be vastly outnumbered by wild-type species due to the abundance of immune cells or the interspersed nature of some tumors throughout normal tissue. This problem is profoundly exaggerated in the case of cell-free DNA/RNA diagnostics, whether from malignant [2, 3], transplant [4], or fetal sources [5, 6], and relies on brute force counting by either sequencing or digital PCR (dPCR) [7] to yield a detectable signal. For these applications, a technique to deplete specific unwanted sequences that is independent of sample preparation protocols and agnostic to measurement technology is highly desired. CRISPR (clustered regularly interspaced short palindromic repeats) and Cas (CRISPR associated) nucleases, such as Cas9, function in bacterial adaptive immune systems to remove incoming phage DNA from the host without harm to the bacteria’s own genome. The CRISPR-Cas9 system has attained widespread adoption as a genome editing technique [8–11]. When coupled with single guide RNAs (sgRNAs) designed against targets of interest, Streptococcus pyogenes Cas9 binds to 3’ NGG protospacer adjacent motif (PAM) sites and produces double-stranded breaks if the sgRNA successfully hybridizes with the adjacent target sequence (Fig. 1a). In vitro, Cas9 may be used to cut DNA directly, in a manner analogous to a conventional restriction enzyme, except that the target sequence (outside of the PAM site) may be programmed at will and massively multiplexed without significant off-target effects. This affords the unique opportunity to target and prevent amplification of undesired sequences, such as those that are generated during next-generation sequencing (NGS) protocols. [Figure: a S. pyogenes Cas9 protein binds specifically to DNA targets that match the ‘NGG’ protospacer adjacent motif (PAM) site. Additional sequence specificity is conferred by a single guide RNA (sgRNA) with a 20 nucleotide hybridization domain. DNA double strand cleavage occurs three nucleotides upstream of the PAM site. b Depletion of Abundant Sequences by Hybridization (DASH) is used to target regions that are present at a disproportionately high copy number in a given next-generation sequencing library following tagmentation or flanking sequencing adaptor placement. Only non-targeted regions that have intact adaptors on both ends of the same molecule are subsequently amplified and represented in the final sequencing library] In this paper, we have exploited the unique properties of Cas9 to selectively deplete unwanted high-abundance sequences from existing RNA-Seq libraries. We refer to this approach as Depletion of Abundant Sequences by Hybridization (DASH). Employing DASH after transposon-mediated fragmentation but prior to the following amplification step (which relies on the presence of adaptor sequences on both ends of the fragment) prevents amplification of the targeted sequences, thus ensuring they are not represented in the final sequencing library (Fig. 1b). We show that this technique preserves the representational integrity of the non-targeted sequences while increasing overall sensitivity in cell line samples and human metagenomic patient samples. Further, we demonstrate the utility of this system in the context of cancer detection, in which depletion of wild-type sequences increases the detection limit for oncogenic mutant sequences. The DASH technique may be used to deplete specific unwanted sequences from existing Illumina sequencing libraries, PCR amplicon libraries, plasmid collections, phage libraries, and virtually any other existing collection of DNA species. Existing specific sequence enrichment techniques — such as pull-down methods [5, 12–14], amplicon-based methods [3, 15], molecular inversion methods [16–18], COLD-PCR [19], competitive allele-specific TaqMan PCR (castPCR) [20], and the classic method of using restriction enzyme digestion on mutant sites [21] — can effectively enrich for targets in sequencing libraries, but these are not useful for discovery of unknown or unpredicted sequences. Brute force counting methods also exist, such as dPCR [3, 7], but they are not easy to multiplex across a large panel. While high-throughput sequencing of select regions can be highly multiplexed to detect rare and novel mutations, and barcoded unique identifiers can overcome sequencing error noise [22], it is costly since the vast majority of the sequencing reads map to non-informative wild-type sequences. A number of sequence-specific RNA depletion methods also currently exist. Illumina’s Ribo-Zero rRNA Removal Kit and Ambion’s GLOBINclear Kit pull rRNAs and globin mRNAs, respectively, out of total RNA samples using sequence-specific oligos conjugated to magnetic beads. RNAse H-based methods, such as New England BioLab’s NEBNext rRNA Depletion Kit similarly mark abundant RNA species with sequence-specific DNA oligos, and then subject them to degradation by RNAse H, which digests RNA/DNA hybrid molecules [23]. These methods are all employed prior to the start of library prep, and are limited to samples containing at least 10 ng to 1 μg of RNA. DASH, in contrast, depletes abundant species after complementary DNA (cDNA) amplification, and thus can be utilized for essentially any amount of input sample. RESULTS We demonstrate deletion of unwanted mitochondrial rRNA using DASH first on HeLa cell line RNA (Fig. 2) and then on CSF RNA from patients with pathogens in their CSF (Fig. 3), in order to increase sequencing bandwidth of useful data. Selection of rRNA sgRNA targets was based on examining coverage plots for standard RNA-Seq experiments on HeLa cells as well as on several patient CSF samples. Coverage of the 12S and 16S mitochondrial rRNA genes was consistently several orders of magnitude higher than the rest of the mitochondrial and non-mitochondrial genes (Figs. 2c and 3). We chose 54 sgRNA target sites within this high-coverage region of the mitochondrial chromosome, situated approximately every 50 bp over a 2.5 kb region (sequences