Genetic data formats: what FASTQ, VCF, PLINK and others are

A transparent 3D DNA helix displaying numeric and letter sequences with highlighted markers, serving as the main cover image for the article on genetic data formats.

When your saliva sample reaches the laboratory, it goes through several stages of processing before becoming a report with results. At each stage, the data are stored in a different format: first as raw sequences, then as an alignment map, then as a table of detected variants, and finally as a structured dataset for statistical analysis. For example, a single indicator in your report on a predisposition to elevated cholesterol is based on data that existed in FASTQ, BAM, VCF and PLINK formats at different stages before being converted into a single score. The most common mistake is to assume that “genetic data” means one file. In reality, it is a pipeline of formats, each corresponding to a specific stage of analysis.

To understand what happens between your saliva sample and your report, it is enough to know four key formats: FASTQ, BAM, VCF and PLINK. Together, they form a standard analytical pipeline in modern genomics. Let’s look at each one in turn.

FASTQ: raw data directly from the sequencer

FASTQ (pronounced “fast-cue”) is the format used to store raw sequencing results. The name comes from the abbreviation FASTA (a format for storing DNA sequences) with the addition of the letter Q from Quality. Imagine a sequencer reading your DNA like text but not being equally certain about every letter: FASTQ stores both the sequence itself and a confidence score for each individual nucleotide.

Each entry in a FASTQ file consists of four lines: the read identifier, the nucleotide sequence itself, a separator, and a line of quality scores. For example:

@read_001

ACGTTAGCATGC...

IIIIHHGFEDCB...

The quality line encodes the sequencer’s confidence in each “letter”: the higher the character on the ASCII scale (American Standard Code for Information Interchange), the more confidently that nucleotide was read. FASTQ files are large, typically several gigabytes per sample, and serve as the starting point for all subsequent analysis. On their own, they are not suitable for interpretation: they are not yet “genetic data” in the sense of specific variants, but rather the raw material.

A central glowing orange sphere with illuminated data fibers converging from all sides through transparent crystals, illustrating population-scale data processing and PRS computation.

BAM: sequences aligned to the genome

The next step after FASTQ is alignment. Software compares each short fragment with the human reference genome and determines where it belongs. The result is stored in SAM format (Sequence Alignment Map) or in its compressed binary version, BAM (Binary Alignment Map). In practice, BAM is generally used because it takes up substantially less space while storing the same data.

A BAM file can be thought of as a huge table: each row describes one DNA read, its position in the genome, orientation, alignment quality and additional technical information. Together, all rows form the genome coverage — how many times each position was read. The higher the coverage, the more reliable variant detection is at the next stage.

BAM files are even larger than FASTQ files and require a special index file, BAI (BAM Index), to quickly locate specific regions without reading the entire file. For most people receiving a genetic report, BAM remains “behind the scenes”: the laboratory processes it automatically and does not provide it to the end user.

VCF: a table of detected variants

After alignment, the algorithm compares covered positions with the reference genome and records the locations where your DNA differs. These differences are stored in VCF format (Variant Call Format). VCF is the first format in which specific genetic variants appear: SNPs (single-nucleotide polymorphisms), small insertions and deletions.

Each row of a VCF file describes one variant: the chromosome and position, the reference nucleotide, the alternative nucleotide in your genome, the call quality and the genotype. For example, a row may show that at position 5246696 on chromosome 11 you have G instead of the reference A — meaning that you carry a particular variant in the HBB gene. The rs numbers you see in the marker table of your report are taken from VCF files.

VCF is a standard exchange format used by laboratories, researchers and analysis software. If you want to upload your own genetic data to an open database such as ClinVar or dbSNP, VCF is the most common input format for this purpose.

A flowing stream of digital code and data passing through a structured geometric canyon of light blocks, illustrating the genetic data processing pipeline.

PLINK: a format for population and polygenic analysis

VCF is convenient for storing variants from one individual but less convenient for analysing thousands of people at once. Population genetics and the calculation of polygenic risk profiles (PRS, Polygenic Risk Score) commonly use the PLINK format, developed by the software of the same name for statistical analysis of GWAS data (Genome-Wide Association Study).

PLINK stores data in three linked files. The .bed file contains genotypes in a compressed binary format. The .bim file describes each SNP: chromosome, position, rs number and both alleles. The .fam file contains information about individuals: their identifiers, sex and, where available, phenotype data. Together, the three files form a single dataset used by PRS calculation algorithms.

Think of PLINK as a large spreadsheet: rows are people, columns are SNP positions, and the values in the cells are genotypes. This is the form in which large population cohorts, such as UK Biobank with nearly 500,000 participants, provide data for GWAS studies, whose results then underpin the polygenic profiles in your report.

Other formats: FASTA, BED and GFF

In addition to the four main pipeline formats, several others are regularly used in genomics:

FASTA (from Fast Alignment, although the name developed historically) is a basic format for storing DNA or protein sequences themselves, without quality information. The human reference genome GRCh38 (Genome Reference Consortium Human Build 38) is stored in FASTA format and is used as the reference against which your data are aligned.
BED (Browser Extensible Data) is a format for describing genomic coordinates: exactly where a particular region starts and ends on a chromosome. It is used to define regions of interest, such as exons in WES (Whole Exome Sequencing).
GFF / GTF (General Feature Format / Gene Transfer Format) are formats used for genome annotation: where genes, exons and regulatory elements are located. When your report states that an SNP is located in or near the BRCA1 gene, that information comes from GFF annotation.

How these formats relate to your Apixmed Prism report

None of the formats described above is delivered to you directly, and that is appropriate. The analytical pipeline from a saliva sample to the report looks like this:

The exact set of formats depends on the analysis method. With whole-genome sequencing (WGS, Whole Genome Sequencing), the pipeline passes through all four formats: FASTQ → BAM → VCF → PLINK. With microarray genotyping, FASTQ and BAM are absent: the microarray produces an intensity table directly, which is converted into VCF, after which the analysis follows the same path. This is why genotyping is faster and less expensive, but does not detect variants outside the chip.

Sequencing or genotyping — output: FASTQ and VCF, or VCF only, respectively
Alignment to the reference genome — result in BAM format
Variant detection — SNP table in VCF format
Polygenic profile calculation — conversion to PLINK and application of weights from GWAS studies
Interpretation and report — the numerical score and position on the scale that you see on each indicator page

The rs numbers you see in the marker table of your report are identifiers for specific rows in the VCF file from your analysis. If you download your own raw data to explore them independently or transfer them to another service, VCF is the most convenient format for this purpose because it is standardised and compatible with most analytical platforms.

For an explanation of what the numerical results in your report mean, how to read the scale, and where the line lies between predisposition and diagnosis, see the article How to read your Apixmed Prism report: what the results mean.

The format determines what can be done with the data

FASTQ captures the raw material: sequences and their quality. BAM shows where those sequences are located in the genome. VCF describes specific differences from the reference. PLINK makes it possible to analyse those differences across large groups of people and calculate polygenic profiles. Each format represents one step from a saliva sample to the result in your report.

Understanding this pipeline changes how the report is perceived: the result is not an algorithm’s “opinion” but the endpoint of a standardised analytical process in which every step is documented and reproducible. The rs numbers in the marker table, the reference genome and an SNP weight are all concrete artefacts from specific files.

If you have not yet taken a test or want to expand your genetic profile, explore the available Apixmed Prism genetic testing options.

Genetic test results are not a diagnosis and do not replace a medical consultation. The Apixmed Prism report provides genetic context that complements examination results and helps you make decisions together with your doctor.

References

1. Cock, P. J. A., Fields, C. J., Goto, N., et al. (2010). The Sanger FASTQ file format for sequences with quality scores, and the Solexa/Illumina FASTQ variants. Nucleic Acids Research, 38(6), 1767–1771. https://doi.org/10.1093/nar/gkp1137

2. Danecek, P., Auton, A., Abecasis, G., et al. (2011). The variant call format and VCFtools. Bioinformatics, 27(15), 2156–2158. https://doi.org/10.1093/bioinformatics/btr330

3. Purcell, S., Neale, B., Todd-Brown, K., et al. (2007). PLINK: A tool set for whole-genome association and population-based linkage analyses. American Journal of Human Genetics, 81(3), 559–575. https://doi.org/10.1086/519795

4. Li, H., Handsaker, B., Wysoker, A., et al. (2009). The Sequence Alignment/Map format and SAMtools. Bioinformatics, 25(16), 2078–2079. https://doi.org/10.1093/bioinformatics/btp352