Capture based assay for MRD detection

Table of Contents

1. Introduction
2. Pipeline summary
3. Pipeline structure
4. References
5. Usage
6. Output
7. Citation
8. Contact

Introduction

 This repository describes a Nextflow pipeline for the analysis of error-corrected sequencing data using fgbio tools. Individual sample libraries incorporate an 8bp Unique molecular index (UMI) tag. These libraries were subjected to target enrichment using a 21-gene panel comprising 192 probes, following the IDT XGen capture protocol.
 Sequencing of these libraries generated three reads per sample: the UMI, the forward read and the reverse read. These reads are given as input to the pipeline in the .fastq.gz format. Read1 is assumed to contain a 8 bp UMI. read2 and read3 being the forward and reverse reads of 151 bases each. The downstream processing steps for these sample reads are mentioned in the following section.

Pipeline summary

flowchart LR

%% Preprocessing
C[Input 
Data] --> D[Add \n UMI]

D --> G[Map & sort \n bam]
G --> G1[Uncollapsed \n bam]

%%Uncollapsed arm
G1 --> S[hsmetrics]
G1 --> U[Coverage]
G1 --> V2[Variant calling
Mutect2, Vardict, Varscan]
V2 --> B3[Variant annotation ANNOVAR]
B3 --> C2[Combine \n variant calling \n data]
C2 --> F[Final Output]
S --> F 
U --> F

%% Consensus alignment
G --> G2[Split bam]
G2 --> K["fgbio tools \n GroupReadsByUmi --> CallMolecularConsensusReads"]
K --> L["Combine bams --> \n Map Consensus bam"]
L --> M["fgbio tools \n Filter Consensus bam"]

M --> Q["sort & index \n collapsed bam"]

%% Metrics
Q --> R[hsmetrics]
Q --> T[Coverage]

%% Variant calling (collapsed)
Q --> V[Variant calling
Mutect2, Vardict, Varscan]

%% Annotation (collapsed)
V --> A1[Variant annotation ANNOVAR]

%% Combine callers
A1 --> C1[Combine \n variant calling \n data]
R --> F[Final Output]
T --> F[Final Output]
C1 --> F[Final Output]

Loading

Pipeline structure

Based on the nfcore pipeline structure, this repository contains:

assets/					# Folder containing reference files
bin/					# Folder with scripts called in the pipeline
modules/				# Folder containing individual process descriptions
sequences/				# Input sequences
mrd_capture.nf			# Nextflow file defining the pipeline
nextflow.config			# File describing input parameters and computing resources for individual processes

References

Execution of this pipeline requires certain reference files. These need to be downloaded and the following parameters need to be modified in the params section of the mrd_capture.config before executing the workflow:

• genome = Complete path to the human genome fasta file(hg19_all.fasta). Please ensure that the BWA index files (hg19_all.fasta.fai, hg19_all.fasta.amb, hg19_all.fasta.ann, hg19_all.fasta.bwt, hg19_all.fasta.pac, hg19_all.fasta.sa) are also present in the same genome folder. The assets folder currently contains placeholder genome and index files.

• annovar_db = Complete path to the humandb database folder for ANNOVAR ( To download additional databases in humandb folder, please refer: https://annovar.openbioinformatics.org/en/latest/user-guide/startup/ ; humandb database used from ANNOVAR version 2020June08)

• bedfile = This file needs to be updated based on the probes used for the assay

• outdir = Location to write the output folder

• gen_ref = Complete path to the gene_fullxref.txt file as downloaded from the ANNOVAR site

Usage

1. Clone the repository using git clone git@github.com:patkarlab/LSC_Capture_MRD.git

2. Enter the directory cd LSC_Capture_MRD

3. Download the reference as mentioned in the Reference section above.

4. Transfer the sample input files *.fastq.gz inside the sequences/ folder.

5. Modify the samplesheet.csv. The sample_ids, without the file extension, should be mentioned in samplesheet in the following format -
   sample1
   sample2
   sample3
   Please remove any empty lines in the samplesheet before running the pipeline.

6. To launch the pipeline, use the following command

nextflow -C mrd_capture.config run mrd_capture.nf -entry MRD_PROBE -bg -profile docker -resume

Output

Samplewise output folders are written to the folder name mentioned in the outdir param in the config file.
Individual output folder contains:

• Samplename_collaps.xlsx : Excel file with annotated variants and the coverage values for the error corrected bam file
• Samplename_cons_sortd.bam : Error corrected bam file (and its index)
• Samplename_collaps_hsmetrics.txt : gatk hsmetrics output for the error corrected bam file
• Samplename_uncollaps.xlsx : Excel file with annotated variants and the coverage values for the uncollapsed bam file
• Samplename_uncollaps.bam : bam file (and its index) before error correction.
• Samplename_uncollaps_hsmetrics.txt : gatk hsmetrics output for the uncollapsed bam file

Citation

If you use this pipeline in your research, please cite:

Leukemic stem cell MRD refines relapse-risk beyond conventional FCM and NGS-based approaches in intensively treated AML. 2026

Contact

Patkarlab