How to Split FASTA and FASTQ Files in OmicsBox

Setting up the tool

The splitter is available in the General tools module of OmicsBox, under FastQ Tools → Split FastQ/FastA files. The wizard accepts one or more files at once and can process a mix of FASTA and FASTQ inputs in the same run, applying the same splitting operation to all of them.

For each file, the user indicates whether it is single-end or paired-end. The paired-end option is only available for FASTQ files and relies on standard suffix conventions to automatically detect read pairs, so that R1 and R2 are kept consistent across the resulting output files (Figure 1).

Splitting options

The tool offers four different splitting modes, selectable from the configuration page of the wizard (Figure 2). The appropriate mode depends on the biological question and the downstream analysis:

• Split By Parts: Divides each input file into N equal parts. Useful when you need to parallelize a downstream analysis across several cores or cluster nodes, or simply to break a very large file into chunks of predictable size.
• Split By Sequences: Splits each input file into multiple files with at most N sequences per output. Ideal when a downstream tool has limits on the number of sequences per run, or when you want uniformly sized batches regardless of the total file size.
• Group By ID: Splits each input file into multiple output files, each containing all the sequences whose ID matches a common pattern. This is particularly useful for multi-species FASTA files, UniProt databases, or any dataset where sequences can be naturally grouped by a shared identifier.
• Extract By ID: Produces a single output file containing only the sequences whose ID matches a user-defined pattern. This is the mode to choose when you want to pull a specific subset of sequences out of a large database.

Splitting sequences by ID

The Group By ID mode relies on regular expressions to decide how sequences should be distributed into files. To make this accessible to users with no regex experience, OmicsBox provides four predefined grouping patterns that cover the most common use cases (Figure 3):

• First word: groups sequences sharing the same first word of the ID (everything up to the first whitespace).
• Until - or _: groups sequences sharing the same leading characters, up to the first dash or underscore.
• Species name: groups sequences by Genus_species prefixes, which is a common convention in comparative genomics datasets.
• UniProt Accession: groups sequences by UniProt accession, regardless of its position in the header.

A custom Regex expression option is also available for advanced users who need finer control. As a rule of thumb, very specific patterns (e.g. matching a single known ID) are better suited for the Extract By ID mode, since every unique match in Group By ID generates a new output file.

Extracting specific sequences

The Extract By ID mode is designed for the opposite use case: instead of partitioning the whole file, it returns a single output containing only the sequences of interest. Users can choose between three patterns (Figure 4):

• Starts with: extracts sequences whose ID starts with a given text.
• Contains: extracts sequences whose ID contains the given text in any position.
• Regex expression: full flexibility for advanced use cases.

Output

The resulting files are saved in the directory specified by the user. They are automatically named using the input file name plus a distinctive suffix, so the provenance of each output is always clear at a glance.

When splitting by parts or by number of sequences, a numeric suffix is added, for example

SRR6312174.part_001.fastq.gz or SRR6312174.part_002.fastq.gz.

When grouping by ID, the matched ID pattern is used directly as the suffix for each file, for example hairpin.part_hsa.fa or hairpin.part_mmu.fa