Alignment Nodes
Alignment nodes map sequencing reads onto a reference. They are typically the most memory- and CPU-intensive step in a pipeline — choose an appropriate resource tier.
BWA-MEM
The standard short-read aligner for DNA (WGS/WES).
Inputs: reads: FASTQ, reference: FASTA
Outputs: aligned: BAM
| Parameter | Type | Default | Description |
|---|---|---|---|
threads | Int | 4 | Alignment threads |
readGroup | String | — | @RG string (required for GATK) |
markShort | Bool | true | Mark shorter split hits secondary |
Bowtie2
Fast, memory-efficient short-read aligner, popular for ChIP-seq and metagenomics.
Inputs: reads: FASTQ, reference: FASTA
Outputs: aligned: BAM
| Parameter | Type | Default | Description |
|---|---|---|---|
threads | Int | 4 | Alignment threads |
preset | String | --sensitive | Sensitivity preset |
STAR
Splice-aware aligner for RNA-seq. Memory-hungry — building the genome index for a human genome needs ~30 GB+ RAM.
Inputs: reads: FASTQ, reference: FASTA, annotation: GTF
Outputs: aligned: BAM, counts: Matrix
| Parameter | Type | Default | Description |
|---|---|---|---|
threads | Int | 8 | Alignment threads |
sjdbOverhang | Int | 100 | readLength - 1 for junction db |
quantMode | Bool | true | Emit gene counts alongside BAM |
Post-alignment helpers
| Node | Purpose |
|---|---|
| Sort & Index BAM | Coordinate-sort and index a BAM (samtools). |
| Mark Duplicates | Flag PCR/optical duplicates (Picard/GATK). |
| BAM → CRAM | Reference-compress alignments to save storage. |
Choosing RAM
| Reference | Suggested tier |
|---|---|
| Bacterial genome | Small (16 GB) |
| Human WES | Medium (64 GB) |
| Human WGS / STAR | Large (128 GB)+ |