Share this page!

Enter URL or ID to Unroll

You can paste full URL like: https://x.com/threadreaderapp/status/1644127596119195649
or just the ID like: 1644127596119195649

How to get URL link on X (Twitter) App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
JWeiland Profile picture
JWeiland
@JPWeiland
Scientist. Infectious disease modeler. Tweets and spelling mistakes are my own Models and analysis seen on @Newsweek @Fortunemagazine @FaceTheNation @NYMag
Perpetual Mind Profile picture Renee Despres, PhD, MPH Profile picture Ekinnajay Profile picture Wayne L Pendley Profile picture AAH Profile picture 13 subscribed
Feb 1 • 7 tweets • 3 min read
A set of 8 highly mutated variant designated BA.2.87.1 were released from South Africa today from Sept-Nov 2023.

While this deserves close watch, at this point, I don't think it's as concerning as the arrival of Pirola. I'll explain my reasoning:

1/
Image It's apparently been 4 months since the first sample collection date, and it still hasnt been detected in any other country that does sequencing.

When Pirola first appeared, there were sequences spotted all over the world within a week or two of first detection.

2/
Jan 24 • 4 tweets • 2 min read
Biobot update 1/23/2024:

Wastewater receding quickly off of the peak, though still near 1 million new infections a day in America. Decline will continue.

🔸990,000 new infections/day
🔸1 in every 330 became infected today
🔸1 in every 33 people currently infected Image My latest model run forecasts a steady decline of infections, with no variants competing for dominance against JN.1.

Jan 16 • 4 tweets • 2 min read
Wastewater update:

Biobot had adjusted their last two weekly estimates down a bit. My Nov. 30th model missed the peak number by 0.6% 😅. Numbers still very high but dropping.

🔸1,270,000 new infections/day
🔸1 in every 26 Americans currently infected
Image This model was made on Nov 30, when our last biobot data point was from 11/22, 5 full weeks out from the peak readings.
Dec 17, 2023 • 4 tweets • 1 min read
In my opinion, 68% XBB is misleading. TODAY it's already less than 50% and dropping like a rock. People probably don't care what the situation was 3 weeks ago.

They want to know what's happening *now* and what is going to happen.

No mention of the fast growth from JN.1. This is very simple to explain. The BA.2.86 descendent JN.1 is quickly outcompeting and displacing the XBB sublineages, and is pushing metrics upwards in many countries.

Why are we dancing around the facts?
Dec 17, 2023 • 5 tweets • 2 min read
This is why we haven't changed Pirola's name for JN.1. This quick evolution was actually predicted by our team as part of Pirola. Even the exact site it needed (S:455) was predicted by members of this team (see @SolidEvidence quote in next tweet from September) BA.2.86 was able to compete equally with any of the fastest lineages out there, even though it was still very weak against class 1 antibodies. Marc Johnson noted the 455 mutation would give Pirola its full potential, before it even evolved.

Dec 15, 2023 • 4 tweets • 2 min read
I can help explain with models why trajectory changes "around 50% prevalence" from a growing escape variant/group and why it's a useful *rule of thumb*.

Plot it on a log chart. 40-60% is where the elbow is. Doesn't guarentee a wave, but a change in trajectory, absolutely.

Image
Image
Image The faster the growth, the *sharper* the elbow. Slower growth? Weak elbow. Multiple variants growing at different speeds? Multiple smaller elbows. JN.1 has enough growth to push numbers up in most countries.
Image
Image
Nov 29, 2023 • 4 tweets • 1 min read
Emergency room visits for Covid seeing a sharp increase in Ile-de-France, where Pirola (JN.1) has recently gained dominance.

We've been expecting this while we watched JN.1 grow. Image We also see new hospital admissions increasing. The Pirola wave has arrived in Ile-de-France.

Note the data source unfortunately only goes back to week 52 of 2022. Image
Nov 18, 2023 • 4 tweets • 2 min read
Why did XBB.1.5 not cause as big of an impact as BQ or the F456L group this summer, though it had a stronger growth advantage? I think we have some answers.

First off, it hit at an ideal time, right after the holiday surge, but I don't think that's the full explanation.

1/3 Image We also have to consider why each variant has growth advantage.

XBB.1.5's "trick" was that it had better ace binding, not that it had better immune escape. So it didn't open up a new pool of susceptibles, it just cleaned up the current pool faster.
Oct 21, 2023 • 6 tweets • 2 min read
We now have updated data from the CDC and Biobot to compare to the forecasts I've put out. This thread are the results vs model, including all my models from earlier this year.

Please evaluate to check my forecast accuracy!

First, the hosp "range" near the peak. Image Next, an "true infections/day" wastewater calibration forecast. Current numbers are almost spot on from my forecast after 3 additional weeks of data. As I expected, the "upswing" in late Sept. was a mirage. Image
Oct 17, 2023 • 5 tweets • 2 min read
#JN.1 #Pirola is demonstrating strong growth in France, where its been found most frequently. At 2% by Oct 2.

These estimates tend to skew high early with a new variant.

To me it looks likely >100%/wk, which would make it more than twice as fast as anything else out there. Image We are also seeing it outcompete the baseline BA.2.86.1 quickly in France. Now more than 50% of Pirola in France are JN.1, in just a few weeks.

If >100% holds, JN.1 could drive a wave in France as early as mid November.

More data can change that timeline. Image
Sep 25, 2023 • 5 tweets • 2 min read
There's a new study claiming no increased risk of new any type infection after covid infection. I don't have a strong opinion on either side of this argument at this point. Let's take an objective look at what this study actually shows (TLDR, it's poor): academic.oup.com/cid/advance-ar… Right off the bat there is a glaring issue. As @DGBassani points out, the control group (uninfected) has more than 3x the deaths as the covid infected group. And there's more, far more comorbidities and +5 years. That's a terrible control group.
Image
Sep 22, 2023 • 4 tweets • 2 min read
BA.2.86 Pirola's global growth advantage is becoming more clear. I excluded South Africa's data (way too messy), and the UK care home cluster.

Estimated weekly advantage is 74%, with a 50% CI of 56-96%.

If this holds, it would suggest Pirola outcompetes the strongest FLips. Image The biggest uncertainty to me still lies in the randomness of sampling. It's unclear it some samples are being sequenced due to SGTF, or more importantly, if that has changed in the past month.

Data pulled from @Mike_Honey_ 's datavis app.powerbi.com/view?r=eyJrIjo…
Sep 14, 2023 • 4 tweets • 2 min read
September 14th US update:

Community spread of Covid still "high", but decreasing after the peak. Current estimates:

🔸650,000 new infections/day
🔸1 in every 510 new people were infected today
🔸1 in every 51 people currently infected Image The decrease is mainly driven by the south, which is seeing a sharp decline in WW. Other areas of the country seeing stabilizing of cases.

Population infection estimates below: Image
Sep 11, 2023 • 5 tweets • 2 min read
Forecast out to mid October:

Prediction is difficult at the moment with a lot of factors at play, so this is a range for now. Peaks are always hard to predict.

Lower model run peaks this week. FLip is growing, but not fast enough to peak later than late Sept. Image FLip, and HV.1 with L452R (last seen in Delta and BA.5*) are growing, with HV.1 and HK.3 being the fastest at the moment. They probably won't come into play until late Oct/Nov, likely starting a new wave.

Still waiting to see how Ba.2.86 Pirola evolves and moves.
Aug 28, 2023 • 4 tweets • 2 min read
One more week of data on hospitalizations to check the model accuracy. Out 5 weeks now.

Model date: Jul 18
Last Hosp data: Jul 8 (always 9-16 days behind current date)
Image Going to post the couple models before that as well. Here's the results from the June 8th model (4-5 weeks):


Image
Aug 20, 2023 • 7 tweets • 2 min read
For sake of comparison, what happened with Omicron? A thread:

First, we had 3 highly mutated sequences from Botswana/South Africa. Within a few days we had enough data from SA to estimate growth advantage of 500%
1/
 Then, we were able to see quick case growth in South Africa. The case growth roughly agreed with the advantage estimate. A couple days after this the Re increased to 3.6.
2/
Aug 17, 2023 • 5 tweets • 2 min read
August 17th update (Biobot): US community spread is back up to "high" with an estimated 610,000 daily new infections.

Similar levels in all 4 US regions.

🔸610,000 new infections/day⬆️
🔸1 in every 550 new people were infected today
🔸1 in every 55 people currently infected Image This rise is still due to EG.5.1, FL.5.1, and XBB.1.16.6 along with some immune waning.

It is not due to the new, heavily mutated variant BA.2.86. There was the US' first sequence reported today (Michigan). We are watching this one *very* closely.

Slide from @jbloom_lab Image
Aug 14, 2023 • 5 tweets • 2 min read
The variants we've seen since 2022 each only have a few S mutations at a time, so their potential was limited.

This unnamed new one has 24.
Omicron had 32.

We don't know yet where this will land on this chart. Given the mutation profile, it has a much larger range of outcomes. Image Maybe our exposure to other Omicrons will stunt it, and maybe it made too many advantage tradeoffs to be highly successful like Omicron initially was. We don't know yet. But we know it's quite different. Here is my count for new S mutations from EG.5.1 (blue also new). Image
Jul 20, 2023 • 6 tweets • 2 min read
Model for next 5 weeks:

Anticipating increases into August, driven by EG.5.1, FL.1.5, and XBB.1.16.6.

Each of these have advantage over XBB.1.16, XBB.2.3 etc.

The lack of cases for the last 4 months is expected to have increased the number of people susceptable to infection. Image It also makes sense that the Northeast is seeing the biggest uptick on wastewater, considering their levels of the faster variants are higher than the rest of the US.

21 day seq data from @RajlabN dashboard
Image
Image
Jul 14, 2023 • 7 tweets • 2 min read
I made a simple Omicron simulation with the following assumptions: R0=8 and avg. protection after infection is 9 months. No seasonal effect etc.

Then I applied NPI that cut the chance of infection per interaction by 7 in 10.

Found the results really interesting!
1/ Image 🔹️3 to 4 waves are expected per year with 0 NPI, even though about 1 infection/yr/capita
🔹️70% effective NPI reduces frequency of waves just less than 2 per year
🔹️But total infections over time were only reduced ~30%
🔹️All the waves started to dampen to flat over time
2/ Image
Jul 9, 2023 • 5 tweets • 2 min read
This data is preliminary (so a lot of uncertainty still), and there isn't a ton of sequencing data, but:

There appears to be a new variant related to XBB.1.9 called FL.1.5.1 that appears to be growing rather quickly in the Dominican Republic. I haven't been interested in any new variant in a good while, but this one might stand out. It appears to have gone from first sequenced case to dominant in less than a month.

I'll be keeping an eye on it and update as needed. More sequencing could show this is just a blip