Tool Parameter Reference

This page documents the hardcoded analysis parameters used by the pipeline's bioinformatics tools. The pipeline supports both prokaryotic and eukaryotic genomes (set via the type field in config.json). The current parameter values are suitable for both organism types under typical sequencing conditions.

---

Short-Read Pipeline

Freebayes (variant calling)

Module: modules/variant_calling.nf

freebayes -f <ref> --min-coverage 10 --min-base-quality 20 --min-mapping-quality 30 --min-alternate-count 3 <bam>

Parameter	Value	Freebayes Default	Description
--min-coverage	10	0	Minimum number of reads covering a locus to call a variant. The default (0) would attempt calls at sites with a single read, producing many false positives. A value of 10 is a standard threshold for reliable variant calling with moderate Illumina coverage.
--min-base-quality	20	0	Minimum per-base Phred quality score. Q20 (99% per-base accuracy) is the widely accepted quality floor for Illumina data. The default (0) would include low-confidence bases, significantly increasing error-driven false calls.
--min-mapping-quality	30	1	Minimum read mapping Phred score. Q30 (99.9% mapping confidence) ensures only confidently placed reads contribute to variant calls. The default (1) would include multi-mapped and ambiguously placed reads, which is problematic in repetitive regions of both prokaryotic and eukaryotic genomes.
--min-alternate-count	3	2	Minimum number of reads supporting the alternate allele. Slightly stricter than the default (2), providing an additional guard against sequencing-error-driven false positives.

Rationale: These are standard community thresholds for calling SNPs and small indels with moderate Illumina coverage (30–100×). Each value departs from the freebayes default to reduce false positives — particularly important in a regulatory/safety context (GMO assessment). The freebayes defaults are intentionally permissive to support diverse use cases (e.g., low-frequency somatic variants); for GMO detection pipelines these permissive defaults would produce excessive noise.

Trade-offs:

• Increasing thresholds improves specificity but may miss low-frequency or low-coverage variants.
• Decreasing thresholds (toward defaults) improves sensitivity but increases the false positive rate.

Delly (structural variant calling)

Module: modules/sv_calling.nf

delly call -g <ref> -o <out.bcf> <bam>

All parameters use Delly defaults. No custom thresholds are applied.

---

Long-Read Pipeline

cuteSV (structural variant calling)

Module: modules/sv_calling.nf

cuteSV <bam> <ref> <out.vcf> <work_dir> -t ${task.cpus}

Parameter	Value	Description
-t	${task.cpus}	Number of threads assigned by Nextflow for this task (bounded by process-level CPU limits in nextflow.config).

All other parameters use cuteSV defaults (minimum SV size 50 bp, minimum support 10 reads, etc.). These defaults are suitable for moderate-coverage PacBio/ONT data on both prokaryotic and eukaryotic genomes.

Sniffles (structural variant calling)

Module: modules/sv_calling.nf

All parameters use Sniffles defaults. No custom thresholds are applied.

DeBreak (structural variant calling)

Module: modules/sv_calling.nf

debreak --bam <bam> -r <ref> -o <out_dir> -t ${task.cpus}

Parameter	Value	Description
-t	${task.cpus}	Number of threads assigned by Nextflow for this task (bounded by process-level CPU limits in nextflow.config).

All other parameters use DeBreak defaults.

---

Reference vs Modified Genome Comparison

NUCmer (whole-genome alignment)

Module: modules/assembly.nf

nucmer --maxmatch -c 100 -b 500 -l 50 <ref> <mod> -p <prefix>

Parameter	Value	Description
--maxmatch	(flag)	Use all anchor matches, not just unique ones. Required to detect duplicated and repetitive regions.
-c	100	Minimum alignment cluster length in bp. Filters spurious short alignments.
-b	500	Maximum gap (bp) between clustered matches. Allows merging across small indels or rearrangements.
-l	50	Minimum exact match (MEM) length in bp. Balances sensitivity with noise reduction.

Rationale: These values are suitable for pairwise comparison of closely related genomes (reference vs. genetically modified organism) for both prokaryotic and eukaryotic organisms. The settings are permissive enough to capture meaningful structural differences while filtering alignment noise.

Trade-offs:

• Lowering -c or -l increases sensitivity to small structural variants but may introduce spurious alignments.
• Increasing -b captures larger structural events but risks merging non-contiguous alignments.

Delta-filter

Module: modules/assembly.nf

delta-filter -m -i 90 -l 100 <delta>

Parameter	Value	Description
-m	(flag)	Keep only the best set of alignments (many-to-many mapping).
-i	90	Minimum alignment identity (%). Filters low-quality alignments.
-l	100	Minimum alignment length (bp). Removes very short, unreliable alignments.

---

Summary Table

Tool	Pipeline	Module	Key Non-Default Parameters
Freebayes	Short-read	variant_calling.nf	--min-coverage 10, --min-base-quality 20, --min-mapping-quality 30, --min-alternate-count 3
Delly	Short-read	sv_calling.nf	Defaults only
cuteSV	Long-read	sv_calling.nf	-t ${task.cpus}
Sniffles	Long-read	sv_calling.nf	Defaults only
DeBreak	Long-read	sv_calling.nf	-t ${task.cpus}
NUCmer	Ref vs Mod	assembly.nf	--maxmatch, -c 100, -b 500, -l 50
Delta-filter	Ref vs Mod	assembly.nf	-m, -i 90, -l 100