"Researchers from the Washington University in St. Louis have developed a new test capable of detecting nearly all viruses known to infect humans and animals. The test could potentially help doctors diagnose infection regardless if they do not have a clue what they are looking for."
...
"What make Virocap different is that physicians do not have to know what they are searching for, said pediatrics professor Gregory Storch."
Seems like it could help with some of those subtle/rare diseases that Dr's have trouble diagnosing.
Can someone with a medical research / science background comment on how significant this is? Real promising or just a sensational press release?
News story
Scientific Abstract
This discussion has been archived.
No new comments can be posted.
Single Test Detects Nearly All Viruses that Infect Humans and Animals
|
Log In/Create an Account
| Top
| 10 comments
| Search Discussion
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
(Score: 2, Interesting) by Anonymous Coward on Friday October 02 2015, @04:10PM
Only detect or also identify?
(Score: 2) by Joe on Friday October 02 2015, @08:21PM
It depends on how you define "detect" vs. "identify".
From the abstract:
"ViroCap could be used to detect viral sequences with up to 58% variation from the references"
The method will detect known viruses and could identify new strains or related viruses of the same family.
- Joe
(Score: 1, Interesting) by Anonymous Coward on Friday October 02 2015, @04:29PM
Sounds like a great diagnostic. But it sounds like a 'great' way to do forensic biometrics. With DNA you can identify genetic family members - take a crime scene blood sample and get a 'partial' match in the FBI's DNA database and you know that the person is probably a relative. With viruses you can do it by association instead of genetics. A partial viral match will tell you that the person lived in close contact with the person on file.
(Score: 1, Interesting) by Anonymous Coward on Friday October 02 2015, @05:42PM
They appear to look at only how sensitive this method is relative to some previous approach. I can't find where they discuss false positive rates.
(Score: 2) by Joe on Friday October 02 2015, @08:32PM
This method amplifies the typically low signal of viral sequences by enriching for known, or closely related, viruses. The false positive rate will depend on where you decide to draw the threshold (the depth and % of the viral genome sequenced), which the research group didn't do.
- Joe
(Score: 0) by Anonymous Coward on Friday October 02 2015, @10:37PM
Well since you can have a method with 100% sensitivity consisting of saying "the virus is there", it makes no sense to talk about that in the absence of specificity. This is well known. I only skimmed the paper, but if they don't address that at all I take that to indicate there is a very high false positive rate they chose not to disclose.
(Score: 2) by physicsmajor on Friday October 02 2015, @05:46PM
Depending on how this test works, the false positive rate has to be astronomical. Most of the population has herpes (HSV-1/2) and latent chicken pox, which are both in the herpesviridae family. There are reports that herpes affects mammals generally.
This study seems to be purely a diagnostic proof of concept. They determined it can detect a lot of stuff in a controlled setting. They even tout the ability for this test to come back positive for viruses in the same family or variations on a theme. Unfortunately, unless they took great pains to exclude these common infections, it will fall flat on its face in the transition to general use on actual people.
(Score: 1, Touché) by Anonymous Coward on Friday October 02 2015, @06:50PM
Wouldn't that be specificity?
(Score: 2) by Joe on Friday October 02 2015, @08:44PM
This research has not developed into a diagnostic test yet, but sequencing can easily differentiate between different strains or related viruses. Signal strength is also dependent on the viral load present in the sample. An active viral infection that is causing disease would typically have many more copies of their genome when compared to one that is latent.
- Joe
(Score: 2) by gringer on Friday October 02 2015, @08:50PM
The innovation of this process is using a shotgun capture approach -- 2 million(ish) probes that you put into a sample mixture and see what sticks. It's a way of excluding unnecessary DNA so that only viral sequences remain for high-throughput sequencing. The actual detection component is not particularly interesting.
Ask me about Sequencing DNA in front of Linus Torvalds [youtube.com]