BINSEQ is a file format for storing nucleotide sequences as binary data. It is designed to be efficient and compact, making it suitable for large-scale genomic data analysis.

To convert a FASTQ (or FASTA) file to BINSEQ we can make use of the bqtools command-line tool. Specifically we can use bqtools encode to convert the file.

Let’s encode a single FASTQ file:

This will run a BINSEQ conversion using default parameters and create a new .vbq file in the current directory.

Let’s compare the file sizes of the original FASTQ file and the resulting BINSEQ file:

You’ll see that that the BINSEQ file is significantly smaller than the original FASTQ file!

By default, bqtools encode will write a VBQ file. This is one of the two variants of BINSEQ and is the more flexible variants that supports variable-length sequences, quality scores, and sequence headers - making it fully lossless to FASTQ.

We can also write a BQ file - the simpler BINSEQ variant that only supports fixed-length sequences without quality scores or headers. This format is more limited than the VBQ format, but can lead to massive throughput improvements for applications that do not require quality scores or headers.

bqtools can auto-determine the mode written using the provided extension (.vbq or .bq)

Let’s compare the file sizes of all the files:

Because BQ does not use any compression (just two-bit encoding of nucleotides), it has a deterministic size from the number of nucleotides in the file.

VBQ uses ZSTD encoding internally and depending on the nucleotide characteristics, the file size is non-deterministic but tends to be smaller than BQ in most cases and roughly the same size as CRAM files.