What actually happens to your spit between Day 1 and Day 60
The first page of a whole-genome sequencing report tells you your ancestry percentages. The fifty pages after it tell you about variants, carriers, and pharmacogenomics. No page tells you what actually happened to the physical tube you mailed in. This is that page.
Day 1 — Collection. You spit into a buffer designed to do two things at once: lyse the cells in your saliva (so DNA comes out of the nuclei) and stabilize the DNA itself for room-temperature shipping. The buffer is why consumer kits don't need dry ice. A kit stored at room temperature for a few weeks will still yield good DNA.
Day 2–7 — Transit. Your tube travels to the lab. For US consumers ordering from a US provider, this is typically 2-4 days. For transatlantic orders (Dante Labs in L'Aquila, YSEQ in Berlin) it can take 7-14 days. This is dead time for the sample and usually the lowest-risk stretch of the whole process.
Day 8 — Extraction. The lab runs your sample through a silica column. The column binds DNA; everything else — proteins, RNA, cell debris, buffer — washes away. A good extraction yields 2–5 µg of intact genomic DNA, which is measured on a fluorometer before any further processing. If the yield is too low, you get re-queued or asked to re-collect.
Day 9–10 — Library preparation. Your DNA is mechanically sheared into fragments around 350 bases long, then chemically ligated to two things: sequencing adapters (so the Illumina machine can grab them) and a sample-specific barcode (so the machine can tell your reads apart from everyone else's on the same flow cell). This step is where human error enters the process; a mis-barcoded library is the most common source of sample-swap incidents.
Day 11–14 — Sequencing. Your library is loaded onto a flow cell in an Illumina NovaSeq (or NovaSeq X Plus, for providers using the newest instruments). Each fragment is amplified into a cluster and read in both directions, 150 bases at a time. A 30× human genome requires roughly 900 million read-pairs. The physical sequencing run is about 48 hours; the queue ahead of your run is typically the longer wait.
Day 15 — Alignment. The raw reads leave the sequencer as a FASTQ file. Software (BWA-MEM2, DRAGEN, or Illumina's own pipeline) maps each read back to a reference human genome — GRCh38 today, or the newer T2T-CHM13 for providers who've migrated. Alignment is compute-bound; a 30× genome runs in 30-60 minutes on modern hardware. The output is a BAM file.
Day 15–16 — Variant calling. The aligned reads are compared to the reference, and the positions where your reads systematically disagree with the reference are written to a VCF file. A typical human genome produces roughly 4-5 million small variants (SNPs and indels), 20,000-30,000 structural variants, and ~50 de novo mutations.
Day 16–60 — Interpretation and reporting. The VCF is joined against clinical databases (ClinVar), population-frequency databases (gnomAD), trait databases, and pharmacogenomic annotations. Only a small fraction of your variants — typically fewer than 0.1% — receive clinically meaningful annotations. Those findings are reviewed (manually, at clinical-grade labs; algorithmically, at consumer-grade labs) and assembled into the report you eventually read.
The bottleneck is not sequencing. It's the interpretation queue. A lab that can turn around a 30× genome in 10 days if sequencing were the only constraint will still take 6-8 weeks because report generation, QC, and clinical review are not automated the way the wet-lab steps are. When a provider advertises a shorter turnaround, they are usually telling you their queue is short. When a provider consistently misses their advertised window, they are usually telling you their queue — and its under-staffing — is the problem.