Abstract
The PSeq data pipeline is designed to transform paired-end short reads from a tagged RNA library into molecule-specific transcript records. It begins by locating the PSeq marker tag, recovering the Source Molecule Identifier (SMID) and strand signal, and grouping reads that originated from one tagged reverse transcript. Within each molecular bin, the pipeline identifies and retrieves the sequence of the source gene, assembles sequence, creates alignment products, and records the evidence chain in a run-specific SQL database.
This paper describes the computational architecture, data products, deployment model, and provenance strategy. It does not present the validation results themselves; the initial library and pipeline evidence is consolidated in Paper 4 of this series.
Key topics
- Marker localization, SMID recovery, and molecule-specific read clustering
- Source-gene identification and reference retrieval
- De novo and reference-guided assembly, alignment, and consensus generation
- Run-specific SQL provenance from FASTQ evidence to molecular records
Contents
- Input model and ten-stage pipeline
- Marker parsing, read trimming, deduplication, and SMID binning
- Source-gene identification and molecule assembly
- Alignment, consensus, and analysis products
- Run-specific SQL record layers and molecular provenance
- Deployment, client reporting, and validation interface