Skip to Main Content
Page banner

Submitted
Being Assessed
In review
Responded
Thanks for your vote
This voting milestone has now ended, and the idea has progressed to a private milestone
Return to idea list
Thanks for approving the idea
This milestone has now ended, and the idea has progressed to a private milestone
Return to idea list

sudo rtPCR Ultra high throughput + Non clinical mass surveillance testing.

E: A connection is helpful
It is pretty clear from the Mega labs that using slow methodical clinical equipment will severely rate limit any purposeful surveillance testing.  Just not enough abi7500's in the country or world. The abi7500's in the Mega labs cant go to 384!. Need to be way above 200,000+ Just hitting 100k is not enough to accurately come out speed up slow down lock down. 

End point pcr is hugely scalable. Kbioscience/LGC processed 1.5 million PCR's per day for gene absence/presence testing.
Currently the NHS/CDC SOP positive Covid criteria is over ct40 and under 40 cycles. But test results do not tell how much Covid you have.
In order to correlate end point to the rtPCR data its possible to use historical rtPCR testing to see if single end point pcr plate read can relate same Covid result. Then you can ignore  the rt curve data and correlate a yes/no Covid result from 35 cycles above 40ct..  

Industrial PCR equipment already exists to do 64 plates pcr in parallel - Hydrocycler. Used for over 20 years in high throughput PCR GSK Sanger CDC..   https://www.selectscience.net/SelectScience-T...s/?videoID=1820

rtPCR very low throughput as one machine = 96 samples per 2 hours  

* So all SOPs stay the same 
* Chemistry (TaqMan) stays the same
* Results stay the same
* Future proof. Ideally need to go to 384 plates to x4 throughput and reduce cost x4.  

***You cant change blocks on 7500's  There are 120 abi7500's !! at Mega Lab. Hydrocycler can do any plate type****

The benefits are hugely increasing throughput and hugely reducing the man power to run 120 abi7500's at Milton Keynes. Easily scalable.

Man power to go into extraction..

___________________________

Non Clinical mass surveillance option

  • NON clinical single point detection at cycle35 on the rtPCR read above ct 40. Throw away the curve data, Patients are not interested in viral load at all. You can correlate Covid Positives data backwards to get confidence in single read read accuracy. 
  • THEN you can use either LGC "Laboratory Gov Chemists" etc who have vast throughput in research genotyping 1.5 million data points per day. or do your self. both Kbioscience historically did 1.5 million data points a day.
  • The data does not go to any patient but used to enable GOV to make decisions on how fast/slow to come out of lock down.. Test whole factory or population. If higher positives, then send swat team to clinical test.
  • Can process x10 x100 times the current throughput 
  • Plus option to pool pre/post rna extraction, NO data goes to any patient, dramatically improves throughput. 
  • Using very high quality proven "research" data is fine for gov surveillance work. 

Jon Curtis & Phil Robinson

How quickly could this be deployed and what are the dependencies?

5 days to correlate historical data and run Hydrocyler on real samples.. Equipment already available to roll out..

What is the likely production volume?

One Hydrocycler can process 64 plates per run in 2 hours. And can run 96, 384 and 1536 pcr plates.. We ran 64-1536 pcr plates 24 hours a day on 6 machines in huge numbers..

96 well plate   =    6,144 data points per 2 hour run
384 well plate  =  24,567 data points per 2 hour run 

 

What are the risks and barriers to using this at scale?

ZERO risks as can correlate data off site, no interference from anyone.

Run off line pre implimentation

Who are you already partnering with on this?

Not been used in sudo rtPCR but zero risk and fast big upside.  

End point surveillance method already running every day. Can be switched to surveillance testing.

edited on Apr 26, 2020 by Jon Curtis
You will need to login to post a comment

Bev Matthews Apr 23, 2020

The idea has been progressed to the next milestone.

Reply 0

Jon Curtis Apr 28, 2020

Also added a complete mass surveillance "Testing the entire population for Covid-19" presentation with roll out, impact, implementation, costs, throughputs, equipment required to this thread.. Its just finished detailing the current application. This is now part of the original submission..

Reply 0

Bev Matthews Apr 23, 2020

Status label added: E

Reply 0

Jon Curtis Apr 28, 2020

Hi Bev we are getting a piece of our pipeline into Milton Keynes Mega lab in the coming days. Difficult to test anything due to their high work load. Even if it reduces their work load.. Whens the next progress deadline..

Reply 0

Jon Curtis Apr 27, 2020

65 million samples processed per month !

Reply 0

Samuel Moses May 25, 2020

Dear Bev, Jon: connecting this to the East Kent Pilot: https://testingmethods.crowdicity.com/post/3296740

NHS Mass PCR Screening Pillar - Pilot for screening all 8000 staff at East Kent: 3CR/KBioScience's end-point PCR as a mass PCR screening covid

Reply 0

Bev Matthews May 26, 2020

Thank you Sam, very helpful

Reply 0

Jon Curtis May 28, 2020

Mass testing data handling capable of 10,000,000 per day 1 person in house.

Reply 0

Jon Curtis May 28, 2020

The ability to data handle 200k-1,000,000 tests per day in rtPCR is well documented to be a considerable task. End point is the reverse, almost zero data handling. Has been tested for 20 years at 1.5 million data points per day with one technician. On Billions of tests >400 genomes and over 1,000,000 various assays.
http://results.lgcgenomics.com/software/krake.../f=overview.htm
KRAKEN LIMS designed specifically for End Point Analysis and processed many Billions of End Point tests since yr2000 Developed by KBioscience now LGC

Users tagged:

Reply 0

Jon Curtis May 28, 2020

Updated full technical file. Tube handling process validated by RNA Automation
https://www.rnaautomation.com largest sample handling company in world. Ironic name though.

Users tagged:

Reply 0

Jon Curtis May 28, 2020

Updated full technical file: Nose to tail. Swab to data.


Tube automation by "RNA-Automation" https://www.rnaautomation.com appropriate name and best in the world at this. RNA Auto are also working on US CDC's tube handling. Their design team have validated our approach as best in class.

Reply 0

Jon Curtis May 29, 2020

Dear all I need an answer to this by Monday thanks..

Hi Sam I’m down your lab with Clare’s and going through the Qnostic SARS cov 2 QC panel. But it does not make any sense to me as this is a test of a whole existing system and does not just test "End Point". Very worrying actually... No design of experiment at all..!!!

The variables pre the "end-point" PCR reaction are huge and if LGC are also testing "end-point" this could result better/worse data only due to the extraction side.
For instance, we could get better looking data tomorrow than LGC, who may actually have better end point chemistry than us - just because the Qnostic samples have been extracted differently.

It is only a useful panel if you are comparing the whole process from start to finish against an alternative which has also been run from start to finish. It does not enable to you to test one individual part of the process, such as the PCR, as it is dependent on the extraction quality of the samples. Just proves little or nothing to be honest.

My strong suggestion based on having optimised commercial mass testing processes for years, is for East Kent to make test plates of known positives/negatives. Then split into multiples so that anyone's end point (or even qPCR) chemistry can be directly compared. The test plate can be a combination of good fresh RNA or poorer quality RNA. Then all data goes back to Claire/yourself blind and for you to score and decide on performance. Twist bioscience make synthetic RNA and you could easily make an exact control dilution plate for all sides to determine LOD CT's Cycling characteristics for all processes.

https://ecommerce.twistdna.com/users/sign_up?...0898.1590695174

Also, as LGC are now carrying out the same tests, would it be good for you to see this and input into their process via Crowdcity as you may want to partner up with them also. I know personally the ability to partner has added considerable value and been its strength. I have offered on many occasions LGC contact yourself and get your real covid samples run in their platforms but offer not taken up. They might be a much better fit than ourselves who knows?

Plus, the best whole process may be a combination of ours and theirs. I have already shared every aspect of ours on Crowd-City site for all to critique plus LGC have had our technical file for weeks. They also need an automation friendly swab/saliva tube and tube handling as well. I also know their extraction side in Berlin are a fab team. Remember every tube in a supermarket has been handled in the same way by the same automation company. No one needs to re-invent this... I would like yourself and Claire to directly compare data from the:

Thermo/Lighthouse End point 2.5ul 5ul 10ul
3CR Mix End point 2.5ul 5ul 10ul
LGC mix End point 2.5ul 5ul 10ul
East Kent clinical End point 2.5ul 5ul 10ul

Correct calling Grouping Distance from NTC NTC movement Cluster tightness Then you can compare same again on extraction chemistry’s and join the best components. Should be a win, win. This gives everyone the best outcome plus can’t be more than a few days to a week's work tops. Plus, if any chemistry is to be used, it is a bloody big thing and would expect more than just a check using this Qnostic plate to be honest.

Phil is very well placed to write a comprehensive design of experiment, as he did this at GSK genetics mass testing labs and KBioscience. He could write then share with all parties. Phil and I are not selling anything so no vested interest. Phil just wants to be able to go down pub and play golf again ASAP. Ta Jon

PS HIV and Ebola are both detected end point, Not exactly new to virology.

Reply 0

Samuel Moses May 29, 2020

Dear Jon

Thanks for your email.
Comparison should bear in level play of exact process mapping I agree.
I have no direct contact with LGC.

Lets see what the Qnostics panel comes back as, albeit the significant reservation that the extraction and other chemistries are not equated to non-confounding.

If that variability affects the results and even if not, a 4 way comparison of different processes cross compared in a 4C3 (combination) manner will help to home in on the most efficient process.

compare data from the:

Thermo/Lighthouse End point 2.5ul 5ul 10ul
3CR Mix End point
2.5ul 5ul 10ul
LGC mix
End point
2.5ul 5ul 10ul
East Kent clinical
End point
2.5ul 5ul 10ul

Correct calling
Grouping
Distance from NTC
NTC movement
Cluster tightness

Perhaps as Crowdicity platform base, and that way all facilitated as National Pathology Alliance Testing Platform aegis.

Reply 0

Share