Biomedical Data Analysis Platform

Home > Analysis Platform > Adapter Trimming and Quality Trimming

Introduction：

Trimmomatic is a widely used, flexible command-line tool designed for preprocessing Illumina Next-Generation Sequencing (NGS) data. Its primary function is to improve downstream analysis accuracy by removing adapter sequences, low-quality bases, and ambiguous nucleotides (N's) from raw FASTQ files. Unlike simple quality trimmers, Trimmomatic employs a "sliding window" approach and supports both single-end and paired-end data. For paired-end reads, it maintains synchronization between read pairs, outputting synchronized files for surviving pairs and separate files for unpaired reads that lost their mate during trimming. It remains a gold standard in bioinformatics pipelines due to its speed, modularity, and reliability.

Workflow Case Analysis Versions Manual Download

A typical Trimmomatic workflow consists of the following sequential steps:

Input Preparation: Provide raw FASTQ file(s) and the appropriate adapter reference file (e.g., TruSeq3-PE.fa).
Adapter Removal (ILLUMINACLIP): Identifies and removes adapter contamination using seed-and-extend alignment against known adapter sequences.
Leading/Trailing Trim: Removes low-quality bases (< threshold) from the start and end of reads.
Sliding Window Trim: Scans the read with a fixed-size window; if the average quality within the window drops below a threshold, the read is cut at that point. This prevents internal low-quality regions from passing filters.
Minimum Length Filter (MINLEN): Discards reads that fall below a specified length after all trimming steps, preventing overly short fragments from entering assembly or mapping.
Output Generation: Writes processed reads to designated output files. For paired-end data, this includes four streams: paired R1, paired R2, unpaired R1, and unpaired R2.

Manual: Key Parameter Reference

ILLUMINACLIP Parameters

These three values control adapter detection sensitivity and clipping behavior in the format ILLUMINACLIP:<adapters.fa>:<seedMismatches>:<palindromeClipThreshold>:<simpleClipThreshold>.

Parameter	Value	Description
seedMismatches	2	Maximum number of mismatches allowed in the seed sequence (typically 16 bp) during initial adapter alignment. Lower values increase specificity; higher values increase sensitivity at the cost of false positives. A value of 2 balances robustness against sequencing errors while avoiding spurious matches.
palindromeClipThreshold	30	Alignment score threshold for "palindrome" mode (paired-end only). In palindrome mode, both R1 and R2 are aligned against each other to detect adapter contamination when the insert size is shorter than the read length. A score ≥ 30 triggers clipping of both reads simultaneously. Higher values reduce false-positive adapter calls in properly sized inserts.
simpleClipThreshold	10	Alignment score threshold for "simple" mode, where each read is independently aligned against the adapter reference. Used when palindrome mode does not apply (single-end data or non-overlapping pairs). A lower threshold than palindrome reflects the reduced confidence of single-read adapter identification.

Quality & Length Trimming Parameters

Parameter	Value	Description
SLIDINGWINDOW (start)	5	Window size (in bases) for the sliding window quality scan. The window begins at position 5 rather than position 1, skipping the first few cycles which often exhibit systematic quality artifacts on Illumina platforms. The average Phred quality within this 5-base window is evaluated at each step along the read.
SLIDINGWINDOW (end)	20	Minimum average Phred quality required within the sliding window. When the running average drops below Q20, the read is truncated at that position. Q20 corresponds to a 1% base-call error rate and is a widely accepted balance between retaining usable data and removing unreliable tails.
LEADING	3	Removes bases from the 5′ end of the read if their individual Phred quality is below Q3. Targets low-quality start positions caused by primer synthesis issues or flow cell edge effects without affecting the rest of the read.
TRAILING	3	Removes bases from the 3′ end of the read if their individual Phred quality is below Q3. Addresses the universal quality decay toward read ends in Illumina sequencing-by-synthesis chemistry.
MINLEN	36	Minimum acceptable read length after all trimming steps. Reads shorter than this value are discarded entirely. This parameter has no universal default; it should be set based on downstream application requirements (e.g., ≥ 36 for variant calling, ≥ 50–75 for de novo assembly, ≥ read length × 0.5 as a general safeguard). Omitting MINLEN risks passing uninformative fragments to aligners or assemblers.