I am wondering if there are techniques for diagnosing pathogenic illnesses with DNA analysis.

For example, imagine a person has a staph infection on their skin. A scraping could be taken and then analysed. In this cases there will be two sources of DNA: the person and the staphylococcus bacteria. My understanding is that statistics can be done to separate the two and thereby positively identify the DNA sequences unique to the staphylococcus, and thus exactly diagnose the pathogen, even telling what family or strain it is.

Now, imagine a harder situation. The patient has a viral infection like a cold. In this case we may still be able to do the same procedure using nasal mucus, but there may be other passive viruses in the mucus. Is it possible?

Now, imagine an even harder situation. The patient has a internal viral infection causing a fever. There blood has thousands of retroviruses in it. Is there any way to isolate the virus causing the fever and get its DNA?


This is all routine, to the point of being trivial. Genome-based diagnosis is very common and is the preferred approach for many, if not most, infectious diseases. Here is a list of over 150 FDA-approved genome-based assays for pathogens.

Some quick examples include influenza:

Rapid Molecular Assays

Rapid molecular assays are a new kind of molecular influenza diagnostic test for upper respiratory tract specimens with high sensitivity and specificity.1 One platform uses isothermal nucleic acid amplification and has high sensitivity and yields results in 15 minutes or less. Another platform uses RT-PCR and has high sensitivity and produces results in approximately 20 minutes.

Other Molecular Assays

Reverse Transcription-Polymerase Chain Reaction (RT-PCR) and other molecular assays can identify the presence of influenza viral RNA in respiratory specimens with very high sensitivity and specificity.

--Influenza Signs and Symptoms and the Role of Laboratory Diagnostics

RSV ...

A newer test protocol (RT-PCR) is more reliable and is replacing the antigen-detecting test in many hospitals and community laboratories.

--Respiratory Syncytial Virus (RSV) Infection

Meningitis ...

Over the past several years, conventional and real-time PCR assays have been developed for detection of bacterial meningitis pathogens. Reliable assays have been extensively evaluated using invasive clinical isolates and/or clinical specimens from around the world (8, 12, 15, 17, 29, 35, 52, 60). In general, validated assays should have high sensitivity and specificity. They can be used as complementary approaches in bacterial disease diagnosis.


and many, many more.

An important trend now is using genome detection methods to look for many different potential pathogens in a single sample; these "lab on a chip" approaches (e.g. here are just entering clinics but will probably be routine in a few more years.

  • $\begingroup$ That is fine if you know what you are looking for, but what if it is uncertain what the disease is? Can you test a sample against all known diseases, or only a few likely guesses? $\endgroup$ – Imprisoned Rhesus Jul 18 '17 at 17:41
  • $\begingroup$ @ImprisonedRhesus - The "shotgun" approach you ask about is unwise and exceedingly expensive. Physicians who do not use their intelligence in making presumptive diagnoses and testing for the most likely culprits first do no one (or no institution) any service. Rare diseases do exist, and are a diagnostic challenge, but to use this approach (test for everything) is just sloppy and calamitously expensive. $\endgroup$ – anongoodnurse Jul 18 '17 at 17:56
  • $\begingroup$ Okay, so you seem to be saying that we can do DNA diagnosis, but only if we have a good idea of what it might be already and can check for that specific possibility, is that right? $\endgroup$ – Imprisoned Rhesus Jul 18 '17 at 18:04
  • $\begingroup$ @Imprisoned Rhesus That's a very different question from your original one, which asks about detecting specific agents. I suggest you ask it as a separate question. Short answer, there are research-level tools for identifying unknown agents, but as far as I know none that are clinically approved $\endgroup$ – iayork Jul 18 '17 at 18:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.