mtb-mixed-infection-pipeline

mtb-mixed-infection-pipeline is a reproducible bioinformatics workflow for detecting mixed Mycobacterium tuberculosis infections from whole-genome sequencing (WGS) data. It integrates Snippy, FreeBayes (pooled-discrete mode) & a patched version of MixInfect2.R to process raw sequencing reads through alignment, joint variant calling, & statistical analysis. This pipeline outputs multi-sample VCF files with key genotype information suitable for identifying mixed-strain infections.

---

---

Reference genome

• Organism: Mycobacterium tuberculosis H37Rv
• GenBank accession: AL123456.3

---

Running Snippy

1. Create conda environment for Snippy

conda env create -f snippy_env.yaml
conda activate snippy-env

2. Run Snippy on multiple samples

sample.list.txt

Format: SampleID<TAB>Forward_Read<TAB>Reverse_Read

TN106985	~/TN106985_1.fastq.gz	~/TN106985_2.fastq.gz
TN106727	~/TN106727_1.fastq.gz	~/TN106727_2.fastq.gz
TN106925	~/TN106925_1.fastq.gz	~/TN106925_2.fastq.gz
TN106439	~/TN106439_1.fastq.gz	~/TN106439_2.fastq.gz

./run_snippy_mtb.sh -i ~/sample.list.txt -r ~/AL123456_MTB_H37Rv.fasta -t 8

---

FreeBayes joint-genotyping (pooled-discrete)

1. Install FreeBayes

conda install -c bioconda freebayes

2. Prepare BAM files

samtools indexed sample bams '**snps.bam**' are available from Snippy run on multiple samples

3. Run FreeBayes

freebayes -f AL123456_MTB_H37Rv.fasta \
  -b TN106985.snps.bam -b TN106727.snps.bam -b TN106925.snps.bam -b TN106439.snps.bam\
  --pooled-discrete \
  --use-best-n-alleles 4 \
  --genotype-qualities \
  --min-alternate-fraction 0.01 \
  --min-alternate-count 2 \
  > multisample.vcf

4. Generate an index of the reference genome

samtools faidx AL123456.fasta

---

Detect mixed tuberculosis infections with MixInfect2

1. Create R environment

conda env create -f r-mixinfect.yaml
conda activate r-mixinfect
conda install -c conda-forge r-data.table r-optparse r-ggplot2 r-mclust "r-base>=4.0" "icu=73.2"

2. Run MixInfect2

This script analyzes multi-strain Mycobacterium tuberculosis infections from VCF files using variant clustering & SNP-based heuristics.

(r-mixinfect):$ Rscript ~/MixInfect2/MixInfect2.R \
  --VCFfile ~/multisample.vcf \
  --prefix output \
  --maskFile ~/MixInfect2/MaskedRegions.csv \
  --minQual 10 \
  --useFilter FALSE

3. MixInfect2 outputs

• output_MixSampleSummary.csv: Summary of sample classifications.
• output_BICvalues.csv: BIC scores & inferred strain counts.

---

Loaded R packages

Loading required package: mclust
Package 'mclust' version 6.1
Type 'citation("mclust")' for citing this R package in publications.
Loading required package: stringr
Loading required package: optparse
Loading required package: foreach
Loading required package: doParallel
Loading required package: iterators
Loading required package: parallel

---

Processing summary

Total variants read from VCF:          590,051
After removing indels:                 576,927
After applying QUAL > 10:                1,121
After removing spanning variants (*):    1,121
After masking regions:                    884

---

Output summary table

SampleName	Mix.Non.mix	hSNPs	Total.SNPs	% hSNPs_totalSNPs	No.strains	Major.strain.prop
TN106985	Mix	459	517	88.78%	2	0.78
TN106727	Mix	477	505	94.46%	2	0.71
TN106925	Non-mix	6	149	4.03%	1	NA
TN106439	Non-mix	9	252	3.57%	1	NA

---

License

This project is licensed under the MIT License.

References

• MixInfect2 GitHub
• Snippy Documentation
• FreeBayes GitHub