Chapter 1 Introduction

IOBRpy workflow

Figure 1.1: IOBRpy workflow

This tutorial presents a reproducible IOBRpy workflow that starts with public paired-end FASTQ files. It includes paired agent skills, Salmon and STAR mode runall, BayesPrism, TME clustering, NMF clustering, and HLA typing.

1.1 Output directories

All chapters use /analysis/IOBRpy_PRJNA1161405_demo as the project root. The complete workflow creates:

/analysis/IOBRpy_PRJNA1161405_demo/
├── input/
│   └── fastq/
├── reference/
│   ├── salmon/gencode44/
│   └── star/
├── results_salmon/
│   ├── 01-qc/
│   ├── 02-salmon/
│   ├── 03-tpm/
│   ├── 04-signatures/
│   ├── 05-tme/
│   ├── 06-LR_cal/
│   └── 07-TCRBCR/
├── results_star/
│   ├── 01-qc/
│   ├── 02-star/
│   ├── 03-tpm/
│   ├── 04-signatures/
│   ├── 05-tme/
│   ├── 06-LR_cal/
│   └── 07-TCRBCR/
├── bayesprism/
├── hla_typing/
├── nmf/
└── tme_cluster.csv

Directory	Contents
`input/fastq/`	The ten paired-end FASTQ samples used in the worked example.
`reference/`	GENCODE v44 reference files and the Salmon and STAR indexes.
`01-qc/`	Cleaned FASTQ files and quality-control reports generated by fastp.
`02-salmon/`	Salmon transcript quantification and merged count and TPM matrices.
`02-star/`	STAR alignments, coordinate-sorted BAM files, gene-level read counts, and the merged count matrix.
`03-tpm/`	TPM converted from the count matrix and the result after log2 transformation.
`04-signatures/`	Sample-level scores for the included biological signatures.
`05-tme/`	Results from the six-method TME panel and their merged profile.
`06-LR_cal/`	Ligand-receptor activity scores calculated from the expression matrix.
`07-TCRBCR/`	TRUST4 T-cell and B-cell receptor repertoire outputs.
`bayesprism/`	BayesPrism immune deconvolution results.
`tme_cluster.csv`	TME clustering results.
`nmf/`	NMF clustering results.
`hla_typing/`	Per-sample HLA typing results and the merged HLA table.

1.2 Example RNA-seq project

The example dataset comes from NCBI BioProject PRJNA1161405.