Thread by @ewanbirney

Right. Deep Breath. RT-PCR "false positives" and Ct numbers (again). tl;dr it is complex, but the RT-PCR testing systems deployed across the world are sound and the people who run them report positives are positives and little can be improved obviously.

Context: I am a genomics/genetics + computational biology expert. I know a large number of infectious disease testing experts. I have a COI in that I am a long established consultant for a company (ONT) that makes a new test here; this gives me additional insight

There a number of classes of false positives which don't concern the current debate (eg, sample swaps, lab contamination). To repeat an early point all the people I know in this are paranoid about this, test and check in a rather detailed way and these are looooow.

The current debate, picked up by the "he built a team that landed rockets on floating ships" Elon Musk is not really "false positive" it is "scope of test".

RT-PCR (and other tests, like LamPORE) test for the presence of SARS_CoV_2 viral RNA in the swab. This is not quite what you want as at the end of infection one can be RNA positive but no infectious viruses.

If one could magic up a quick, sensitive and specific, send-your-swab-in test for infection viruses present that would be great. Sadly noone has done this (yet) (I will come back to lateral flow tests).

RT-PCR (+ other nucleic acid tests) there are not quite the ideal scope, but they do have really good sensitivity and specificity - in the 99% for good samples.

A note now on sensitivity. It is going to be tempting in a moment to use an internal metric on the test (Ct numbers) to change the scope of the test. As well as the inherent complexity of this the big elephant in the room is swab/sample/process variance.

People don't swab well. Swabs are drier or wetter or contaminated with RNases (which are everywhere). Sample shipping can be disruptive and late. Lab processes vary from the profound (missing a sample shipment) to subtle (the thickness of the foil on the reagent cover varies)

All these things means one get substantial additional variance in the real world from more controlled settings. Welcome to ... real life practical infection biology!

Given this variance, and the danger of false negatives in majority of settings, one has far less scope to try to draw boundaries in the Ct space.

The next, massive elephant in the room (it's a crowded room of elephants) is that there are different RT-PCR machines (this is good - diversity of supply) with different RT-PCR locations on the genome looked at, and all these things systematically vary on (for example) Ct values

This is why individual workflows (these primers on these machines with this swab) are regulated and approved and although if a scheme works on RT-PCR on one machine it is likely to work elsewhere... it is not like there is one "Ct" scale.

ok, a reminder on Ct number. Ct number means cycle threshold and it is the number of cycles needed until there is a large increase in flourescence/optical properties on a RT-PCR machine. With the same machine and primers and experiment, Ct numbers are meaningful >>

A low Ct number means more starting RNA. A high Ct number means less starting RNA. One can semi-quantify this to absolute standards with spike-ins if desired.

The suggestion is that one could change the "scope" of the RT-PCR test by shifting the threshold of this Ct number (even cleverer perhaps do this with different Ct space different parts of the viral genome which amplify due to the viral infection cycle: an n-dimensional plane)

It does, as a pencil-and-paper exercise, look sort of attractive. But - the two elephants in the room - real world variation in swabbed samples (let me repeat - there is RNase in all sorts of places) and variation between machines means this is ... not going to work easily.

Furthermore although more RNA usually means more intact virus it is clear that there is biological variation infection (location, size etc) so "amount RNA" does not have a tight relationship to "will this person infect another person".

(I get Elon Musk built a team that landed rockets on floating ships, so "not going to work" is a red rag here. Let's put this another way - it is not going to be a simple thing to work; getting a robust -70 cold chain world wide would be a better use of engineering innovation).

What about Lateral Flow tests? These are interesting because they test for the folded viral protein (antigen) not the nucleic acid. Again - same gotchas apply - swab variation, weird stuff happening and does viral protein equate to infectious virus?

As it happens, the limit of these tests seem to capture low Ct (high RNA levels) individuals, less so high Ct. It is super super hard to do the experiment you want to (will this person transmit to other humans) but at least in vitro the trend is this way.

For sure, if one could make a *more* sensitive Lateral Flow test and/or say portable nucleic acid tests (my COI comes into play here on LamPORE) you would again use all that sensitivity on getting people who are possible infectious identified.