scarecrow is a scRNA-seq pre-processing tool developed to increase the usable data generated from an experiment by idenifying jittered barcodes, correcting mismatches, and outputting the data in a format used by popular third-party tools. Below is a schematic diagram illustrating the scarecrow workflow. This begins with raw FASTQ files and barcode whitelists being passed to seed. The resulting barcode profiles are passed to harvest which outputs a file of expected barcode positions. These positions, together with the raw FASTQ files and barcode whitelists are passed to reap, which returns either a SAM or FASTQ file containing the target sequence and barcode information. Optional steps include extracting a barcode from the FASTQ read header with weed, if the reads were demultiplexed, and removing any reads containing unmatched barcodes with sift.
We recommend adapter trimming after running scarecrow to ensure that read lengths are consistent when pre-processing the data, otherwise there is the potential for trimming to result in offset barcodes that might not subsequently be matched.
If the aim is to perform cell-gene quantification via pseudoalignment using kallisto-bustools, then an interleaved FASTQ file should be generated with scarecrow reap. The interleaved FASTQ contains corrected barcodes followed by the UMI in read 1 and the extracted sequence in read 2, and is supplemented by a JSON file that contains the kb count parameters required to process the FASTQ file. If it is necessary to add a barcode from the FASTQ header, then a SAM file should be generated with scarecrow reap and annotated with scarecrow weed. The SAM file can be converted to an interleaved FASTQ file for use with kallisto by scarecrow recast.
If the aim is to perform cell-gene quantification via standard alignment, for example the UMI-tools workflow, then a SAM file should be generated with scarecrow reap. The SAM file can be aligned with STAR, annoated with featureCounts, and processed with UMI-tools (taking advantage of the read tags in the SAM file) to generate a counts matrix.