COVID tests detect specific pieces of the SARS-CoV-2 virus, not “COVID” as a general illness. The exact target depends on the type of test: some look for the virus’s genetic material, others look for proteins on the virus’s surface, and a third type checks whether your immune system has produced antibodies in response to the virus. Each approach answers a slightly different question, and understanding the differences helps you interpret your results correctly.
What Rapid Antigen Tests Detect
The home tests most people are familiar with are rapid antigen tests. These look for a specific protein called the nucleocapsid protein, which sits inside the virus and helps package its genetic material. When you swab your nose, you’re collecting a mix of mucus, shed cells, and (if you’re infected) viral particles. The test strip contains antibodies designed to latch onto that nucleocapsid protein. If enough viral protein is present in your sample, the antibodies bind to it and trigger a visible line on the test card.
Nearly all commercially authorized antigen tests target the nucleocapsid protein rather than the spike protein that sticks out from the virus’s surface. Of the first 13 antigen kits authorized by the FDA, 12 targeted the nucleocapsid. Some researchers have explored spike-based antigen tests, but nucleocapsid remains the standard because the protein is abundant and relatively stable across variants.
The tradeoff with antigen tests is sensitivity. They need a fairly high concentration of virus in your sample to turn positive. On the first day symptoms appear, rapid tests catch roughly 60% of infections in people with symptoms. For people without symptoms, that number drops to about 9% on the day infection begins. This is why serial testing matters: testing two or three times at 48-hour intervals raises detection in asymptomatic people from about 34% with a single test to nearly 69% with three tests over the first week of infection.
What PCR Tests Detect
PCR tests (formally called nucleic acid amplification tests) look for the virus’s genetic code, its RNA. A lab extracts RNA from your sample, converts it to DNA, and then repeatedly copies tiny target segments of the viral genome. Each round of copying is called a cycle. If viral RNA is present, the copied segments accumulate until they cross a detectable threshold. If no virus is there, nothing gets amplified.
The specific genetic targets vary by manufacturer, but three regions recommended by the WHO are most common: the N gene (which codes for the nucleocapsid protein), the E gene (which codes for the envelope protein), and the RdRp gene (which codes for the enzyme the virus uses to replicate itself). Many commercial test kits target two or more of these genes simultaneously. This multi-target design is a built-in safety net. If a mutation in one gene region prevents detection, the other target still flags the virus.
The result hinges on something called the cycle threshold, or Ct value. This is the number of copying cycles needed before the virus becomes detectable. A low Ct value (under 30) means the sample started with a large amount of virus, so it was easy to find. The standard cutoff for a negative result is 40 cycles. Results between 38 and 40 are generally considered inconclusive. Different test manufacturers may set slightly different cutoffs based on their test design.
PCR tests are the most sensitive diagnostic tool available. At peak viral shedding, about three days after symptom onset, PCR detects roughly 83% of infections compared to 59% for antigen tests on the same day.
What Antibody Tests Detect
Antibody tests are fundamentally different from PCR and antigen tests. They don’t look for the virus at all. Instead, they check your blood for immune proteins (antibodies) your body made in response to a past infection or vaccination. Because of this, antibody tests cannot tell you whether you’re currently infected.
These tests measure two main classes of antibody: IgM, which appears relatively early after infection and fades within weeks to months, and IgG, which builds more slowly but persists much longer. Finding IgM can suggest a relatively recent infection, while IgG alone may indicate an older one.
What makes antibody testing especially useful is that it can distinguish between immunity from vaccination and immunity from natural infection. COVID vaccines train your immune system to recognize the spike protein. If an antibody test finds antibodies against the spike protein, that could mean you were vaccinated, previously infected, or both. But antibodies against the nucleocapsid protein only come from actual infection, since no widely used vaccine contains that protein. So a test targeting nucleocapsid antibodies specifically can tell whether someone has been infected regardless of their vaccination history.
How Variants Affect Test Accuracy
As SARS-CoV-2 mutates, the question of whether tests still work keeps coming up. The answer depends on which part of the virus the test targets. Molecular tests that rely on a single genetic region can lose sensitivity if a mutation changes that region enough. The Omicron variant, for example, carried mutations that caused certain N-gene and S-gene targets to “drop out,” meaning those specific genetic segments failed to amplify during the test.
Tests designed to detect multiple gene targets handle this well. If one target drops out, the others still catch the virus, so overall sensitivity stays intact. In fact, this dropout pattern became a useful early signal: when labs noticed one gene target failing while others succeeded, it flagged the possible presence of Omicron before full genetic sequencing could confirm it. The FDA actively monitors authorized tests against circulating variants and updates guidance when a mutation threatens to reduce a test’s reliability.
When Each Test Works Best
Viral load follows a predictable curve after infection. It rises quickly, peaks around two to three days after symptoms start, and then gradually declines. The highest percentage of positive antigen test results (59% overall, rising to 80% in people with fever) occurs about three days after symptom onset. PCR peaks at the same point but catches more infections because of its greater sensitivity.
If you’ve been exposed but feel fine, a single rapid test taken too early can easily miss the virus. Testing two to three days after a known exposure, then repeating 48 hours later if the first test is negative, significantly improves your chances of catching an infection. For people with symptoms, especially fever or cough, rapid tests perform much closer to PCR’s accuracy when taken a few days into illness.
Sample quality also matters. Labs sometimes run a secondary check to confirm that a swab actually collected enough cellular material to be reliable. A poorly collected sample, one that barely grazes the inside of the nostril, may not pick up enough virus to trigger a positive result even when infection is present. Swabbing deeply and rotating for the recommended time gives the test the best chance of working as designed.