Can You Get Your Genome Sequenced for Free?
Sometimes, but the word “free” is doing heavy lifting. Almost every consumer offer advertised as free genome sequencing is one of three things: a genotyping array that reads under a million positions, a low-coverage sequencing run that is statistically filled in against a reference panel, or a free analysis layer on top of data you already paid for. Genuine 30x whole-genome sequencing at no cost to you comes from two places: a research study you qualify for, or a clinically indicated test ordered by a physician and billed to insurance. Both have conditions attached, and neither reliably hands you a FASTQ file.
This page is about how to tell the tiers apart, what each one can and cannot answer, and how to check the quality of whatever you end up holding.
What the free offers are reading
A SNP array, sold by 23andMe and AncestryDNA and re-analyzed for free by sites like Genetic Genie and Genomelink, hybridizes your DNA to a chip with roughly 600,000 to 900,000 probes. That is about 0.02 percent of the 3.1 billion base pairs in a haploid human genome. The output is a tab-delimited text file with rsID, chromosome, position, and a two-letter genotype. It contains no indels worth trusting and no copy number calls. There are no structural variants, and no ability to see a variant that was not on the chip.
Low-pass whole genome sequencing is the second tier. Here the library is real sequencing, but at 0.4x to 1x mean depth, meaning most positions are covered by zero or one read. The pipeline then imputes genotypes using a haplotype reference panel. Common variants come out well. Rare variants do not come out well, and those are the ones that carry most clinical weight. Imputation borrows information from haplotypes shared across many people, and a variant present in one person in fifty thousand has no such support.
The third tier is 30x short-read WGS. Each base is covered by about thirty reads on average. That is enough for confident diploid genotyping, reliable small indel calls, and workable copy number and structural variant inference. This is what people mean when they say “my genome.” It costs real money to produce, and nobody gives it away without a reason.
Which free routes return data you can work with
Research studies are the most realistic path. Large population cohorts recruit participants, sequence them, and in some cases return a curated report of medically relevant findings alongside ancestry results. Eligibility is usually geographic, and the return is typically a PDF report rather than a CRAM file. If your goal is raw data you can pipe into your own tools, read the return-of-results policy before you consent. Do not leave it until after.
Clinically indicated sequencing is the other path, and it is not self-service. Diagnostic WGS gets ordered when there is a specific question: an undiagnosed condition with features suggesting a genetic cause, a family history that meets testing criteria, or a critically ill infant. The ordering clinician decides, the lab reports on a defined gene set, and you can usually request the raw data afterward under your access rights. Whether that request produces a BAM or a CRAM or a filtered VCF varies by laboratory.
The consumer offers that advertise free sequencing generally fund it by some other mechanism. Read the terms for the data license, the subscription that follows the first year, and whether “whole genome” in the marketing copy means 30x or 0.4x. If the page does not state mean coverage and read length, assume low-pass.
What each tier can answer
The practical question is which analyses your file supports. Array data is adequate for ancestry, for common-variant polygenic scores, and for pharmacogenomic star alleles that happen to be on the chip. It is not adequate for screening rare variants in cancer risk genes, because the chip carries a small selected subset of known pathogenic sites and misses everything else. Consumer array reports of BRCA1 and BRCA2 status are a well-known example of this gap.
Low-pass WGS extends the common-variant picture and adds usable mitochondrial and rough copy number signal. It remains poor for the rare end of the spectrum.
Thirty-fold WGS supports rare variant discovery, indel calling, and mitochondrial heteroplasmy. It also supports CNV calling and repeat expansion genotyping with tools like ExpansionHunter. Even then, coverage is not uniform across the genome. GC-rich first exons and regions with close paralogs drop out. The choice of library preparation and fragmentation method measurably changes how evenly clinically relevant genes are covered.1 Genes with pseudogene homology such as PMS2, SMN1, and CYP21A2 are the usual offenders for short reads.
Interpretation is the harder half of the problem. A typical genome carries millions of variants relative to the reference, and the great majority of the ones that look interesting will be classified as variants of uncertain significance.2 That classification means the evidence does not currently support calling the variant benign or pathogenic, and acting on it is not appropriate. The label is not a placeholder for a future answer you will get next year. Anything that touches medical decision making belongs with a clinical geneticist or genetic counselor working from a CLIA-certified laboratory report.
QC the moment you get files
Before any analysis, establish what you are holding. Start with the reference build, because GRCh37 and GRCh38 coordinates differ and silently mixing them produces wrong annotations rather than errors.
samtools view -H sample.cram | grep '@SQ' | head -3
bcftools view -h sample.vcf.gz | grep -i 'reference\|contig=<ID=chr1,'
A chr1 length of 249,250,621 is GRCh37. A length of 248,956,422 is GRCh38.
Next, measure coverage yourself rather than trusting the vendor summary.
mosdepth --by 1000 --fast-mode --no-per-base sample sample.cram
Look at mean depth and at the fraction of the callable genome above 10x. For a 30x genome we expect roughly 90 percent or more of the genome at 10x or better. If a panel of genes matters to you, pass a BED file to --by and inspect per-exon depth directly.
Then check for sample contamination and for sex and relatedness consistency.
verifybamid2 --SVDPrefix resource/1000g.phase3.100k.b38.vcf.gz.dat \
--Reference GRCh38.fa --BamFile sample.cram
bcftools +guess-ploidy -g b38 sample.vcf.gz
A FREEMIX estimate above about 0.02 means meaningful cross-sample contamination and should prompt a conversation with the lab. Finally, annotate with Ensembl VEP and filter, rather than reading a vendor report.
vep -i sample.vcf.gz --cache --assembly GRCh38 --everything \
--plugin CADD,whole_genome_SNVs.tsv.gz --vcf -o annotated.vcf.gz
If all you have is array data, you can still convert it to VCF for consistent tooling, keeping in mind that the file only contains the probed sites.
bcftools convert --tsv2vcf genome.txt -f GRCh37.fa -s myID -Oz -o array.vcf.gz
What we would do
If the goal is curiosity and ancestry, the free array-upload tools are fine and cost nothing beyond the original kit. If the goal is a durable molecular baseline you will query for years, pay for 30x WGS and insist on the CRAM or FASTQ. Do not settle for a report. Sequencing your genome once at adequate depth beats sequencing it three times badly, since the DNA does not change and reanalysis is cheap compared with resequencing.
It helps to remember how recently this was out of reach at all. The human genome project was conceived in the late 1980s as a decade-scale international effort.3 A commercial sequencing industry existed by 2000, priced far beyond individual reach.4 The cost curve since then is the reason the question on this page is even sensible.
Questions people also ask
Where can I get free whole genome sequencing? Research cohort studies and clinically indicated diagnostic testing are the two routes that produce real 30x data at no cost to you. Both have eligibility requirements, and research studies often return a report rather than raw files.
Can I get my own genome sequenced? Yes. Direct-to-consumer 30x WGS is widely available and does not require a physician order in most jurisdictions. Ask the provider for mean coverage, read length, aligner, variant caller, and reference build before ordering.
Is Genomelink really free? Genomelink’s upload and basic trait reports are free, but the service interprets a file you already own from an array test. It does not sequence your genome, and the reports are limited to the sites on the original chip.
Will insurance pay for genome sequencing? Sometimes, when a clinician documents a medical indication that meets the payer’s criteria. Insurers generally do not cover sequencing requested for general interest or optimization. Coverage decisions and the resulting report belong to the clinical pathway, so start that conversation with a physician or genetic counselor.
Why does my free report miss variants my friend’s WGS found? Because an array only reads the positions its probes target. Any variant not on the chip is invisible, and no amount of reanalysis recovers it.
Footnotes
-
Vanessa Process, Madana M.R. Ambavaram, Sameer Vasantgadkar, et al. Optimization of DNA Fragmentation Techniques to Maximize Coverage Uniformity of Clinically Relevant Genes Using Whole Genome Sequencing. Diagnostics, 2025. https://doi.org/10.3390/diagnostics15182294 ↩
-
A.J. Marian. Sequencing Your Genome: What Does it Mean?. Methodist DeBakey Cardiovascular Journal, 2014. https://doi.org/10.14797/mdcj-10-1-3 ↩
-
P.S Harper. Mapping and sequencing the human genome. Endeavour, 1989. https://doi.org/10.1016/0160-9327(89)90035-5 ↩
-
Kathryn Brown. The Human Genome Business Today. Scientific American, 2000. https://doi.org/10.1038/scientificamerican0700-50 ↩
Pothos builds longitudinal molecular profiles of individuals: whole-genome sequencing, single-cell RNA sequencing, and plasma proteomics, delivered as a detailed report, every data file, and stored cells. Build your profile.