Targeted ExSeq -- Probe Generation v2
Year: 2021
Type: preprint
Abstract: This protocol accompanies Expansion Sequencing (ExSeq), and describes the process of generating padlock probes for targeted ExSeq. The steps detailed here are a generalization of the protocol used to generate padlock probes for figures 4-6, and represent our recommendations for future users of the technology. This protocol has associated code (available at https://github.com/anusinha/TargetedExSeqProbeGen). We have added additional technical explanations here. As a technical summary of the experimental steps of the method, targeted ExSeq begins with physical expansion of a sample of interest (using the chemistry of expansion microscopy), resulting in a tissue-hydrogel composite with the RNA molecules covalently anchored to the hydrogel. An in situ sequencing library is then prepared from a set of transcripts of interest. The library preparation begins with the hybridization of oligonucleotide padlock probes bearing barcodes to RNA transcripts of interest. Multiple padlock probes target one gene, and each of those padlocks have the same barcode identifying that gene. Each gene has a unique barcode. The padlock probes are subsequently enzymatically ligated and amplified using rolling circle amplification (RCA), producing amplicons (often called RCA colonies, or "rolonies" colloquially). The amplified barcode sequences on the amplicon are read out through the process of in situ sequencing, in which sequencing reagents are deployed within the gel. Through iterative rounds of chemistry and imaging, the barcode sequence is read out, one base at a time. The imaging data is processed through the ExSeq image processing pipeline (https://github.com/dgoodwin208/ExSeqProcessing/wiki), resulting in a map of reads in 3D space. This protocol describes the process by which padlock probes are generated. Because the experimental design governs the design of the padlock probes, this protocol is necessarily less prescriptive than the other protocols. The overall architecture of the padlock probes is shown in Fig. 1A (top). The probes are doubly barcoded, enabling the user to read out the barcode sequence either via SOLiD or Illumina sequencing chemistry. The probes consist of the following sequence regions: (1) a split homology region (H1, H2), which hybridizes to the transcript of interest; (2) a barcode for SOLiD sequence readout that is 5' to the Seq Primer; (3) a barcode for Illumina sequencing that is 3' to the Seq primer; and (4) a universal RCA/Seq primer binding site. After RCA, the amplicon concatamer Fig. 1A (bottom) is produced. Note that the concatamer is the reverse complement of the probe sequence. In situ sequencing is then performed on the concatamer. The padlock probe generation process consists of five stages: (1) Target selection; (2) Barcode generation; (3) Barcode assignment; (4) Homology region generation; (5) Probe assembly and export. The first stage of the design is target selection of genes of interest. As this is dependent on the scientific question, generalizable advice is difficult to provide. This protocol assumes that the user has already selected a set of transcripts of interest. Relevant considerations are provided in Step 2. The second and third stages of the design process are closely linked. In these steps, the barcode space is defined (Steps 3-5), and those sequences are assigned to transcripts (Steps 6-7). In the fourth stage, the homology regions of the padlock probes that hybridize to the RNA are identified (Steps 8-10). In the fifth stage, the homology regions, barcode sequences, and backbone sequences are assembled to form the final probe (Steps 11-14). The probes are then ordered and pooled together to form the pooled padlock probe libraries (Steps 15-16), used in the Targeted ExSeq Library Preparation protocol. To enable multiple sequencing chemistries to be used, barcodes are designed in a logical space, which is agnostic to the sequencing chemistry. At the time of assembly, these logical barcodes are converted to nucleotide sequences, which are read out during in situ sequencing with the SOLiD or Illumina sequencing chemistries. The logical barcodes are a repeated sequence of the numbers {0, 1, 2, 3}. For example, 2012310 is a logical barcode of length 7, in which the first base of the barcode is 2, the second is 0, the third is 1, and so on. The mapping between logical barcodes and nucleotide sequences is shown in Fig. 1B. The details of this mapping are discussed in further detail in Steps 11-14. Examples of these mappings being used in the in situ sequencing process are shown in Fig. 1D (Illumina sequencing) and Fig. 1E (SOLiD sequencing). To constrain this design space for this protocol, we focus on a specific implementation of the padlock probe, shown in Fig. 1C. First, the barcode design: we focus on barcodes of length 7, with Hamming distance 3, which enables 1-bit error correction, and 2-bit error detection. This barcode library enables the interrogation of ~330 targets with 7 rounds of sequencing. We also discuss a subset of this barcode library, which enables the interrogation of up to ~40-50 targets with four rounds of sequencing. This subset is of length 4, with Hamming distance 2, enabling 1-bit error detection. The embedded subset is useful for smaller experiments, with 10-50 targets, where the readout can be performed with four rounds of sequencing. Notably, this subset of barcodes is extensible, i.e. additional genes can be added to the library if 7 rounds of sequencing are performed. Second: we design doubly barcoded probes. These padlock probes carry two barcode sequences for readout using either the Illumina or SOLiD sequencing chemistries. The implementation details on the mapping between logical barcodes and the specific nucleotide sequences are shown in Fig. 1F (Illumina) and Fig. 1G (SOLiD). Third: the total probe length is 74 nucleotides, corresponding to the design criterion of a padlock probe being at least 2*n+10 nucleotides long, where n is the total length of the homology region (in this case, n=32 nt). Fourth: we generate up to 16 probes per transcript, which we find strikes a good balance between yield and cost. This protocol was used (with an earlier version of the codebase) to design probes to profile human metastatic breast cancer biopsies as a part of the Human Tumor Atlas Pilot Project (HTAPP). For the design of the SOLiD sequencing barcodes, the alternative sequential barcode design was utilized (see Step 12.2). Worked Example Step 17 is a worked example of generating padlock probes for AKT1, CDK7, and KRT17, using pre-prepared input files. The output results of the script can be compared to reference results.
Authors: Anubhav Sinha, Fei Chen, Shahar Alon, Ed Boyden
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W4240754492