Lots of places, actually.

Who can tell me about snakes?

We monitor WW for Arenaviruses, because they can be scary.

On several occasions we've detected Reptarenavirus sequences, sometimes a lot of it.

They only seem to infect boa constrictors and pythons.

How's it getting in WW?

Yes, sequencing levels are wildly different across the country.

Yes, but they usually spread eventually. BA.3.2 is just exceptionally slow at it.

I hate to say I told you so but...

Since I posted this (3 days ago) there have been over 130 BA.3.2 sequences released (more than doubling the US total).

Every one of them was from NY, MD, NJ, and MA

Shit don't lie.

Data plot of COVID mutations from wastewater.

It's been consistently positive for months, but very slow.

If you look patient samples collect in the last month from those places, these are the numbers that are BA.3.2:

NYC - 13/36 (10/13 kids)
CT - 0/0
PA - 0/0
MD - 7/56 (3/7 kids)

So, it's just uneven sequencing. We will see more patient cases.

2/2

BA.3.2 update.

In the last week BA.3.2 showed up in over 70 wastewater samples, but they were not evenly distributed.

Over half of them came for NY (all NYC), CT, MD and PA.

In NY, PA and CT it was almost every sample (29/31)
In MD it was just over half.

1/

Variant dashboard.

It's a little squishy, coverage is uneven. If you go to our dashboard it shows the fraction of individual mutations from the last 3 weeks (compared to the 3 weeks before). All of the BA.3.2 mutation are increasing in frequency. The average is over 10%, I think.
lung.fish/wastewater-d...

Sequence readout.

This is an example of our readout. We screen sequences for ones that have 2 BA.3.2-specific PMs. The output shows the full genotype of the sequences (to confirm full context), and the fraction of sequences in that region that contained those mutations.

Several groups do sequencing from wastewater, but the output is a mixture of all the lineages. There are various ways to analyze the data to see what all lineages are present. Using our screen (which is pretty stringent) at least half of the samples had BA.3.2 included.

Image of the lack of Marine traffic in the straight of Hormuz.

I'm no sailor, but this doesn't look open to me.

This is weird.

In the last week of US wastewater sampling BA.3.2 was in half of the samples.

But in the same period it was only in 1% of patient sequences.

Maybe this is good; maybe it's so mild they aren't testing.

Or maybe they aren't testing kids.

Yeah, could be.

So you used surrogates for known lineages with the deletion. Is there a way to search for the deletion itself?

Don’t make me block you

Is BA.3.2 Disproportionately Infecting Children? A Look at the Age Distribution Data - Lungfish BA.3.2 sequences show a striking enrichment of children compared to other circulating SARS-CoV-2 lineages across five countries. We investigate why.

Summary about BA.3.2 disproportionately infecting kids. My first ever blog.

lungfish-science.github.io/blog/2026-04...

Wastewater testing finds dangerous nitazenes in St. Louis area schools A dangerous opioid – 5 to 10 times more potent than fentanyl – was found in wastewater from two St. Louis schools. Few St. Louis adults had ever heard of nitazenes, a type of synthetic ...

There are 3 things left that most Americans have a favorable opinion of: rockets to the moon, Dolly Parton, and wastewater testing.

hcrl.wustl.edu/wastewater-t...

How do you search for them?

@ryanhisner.bsky.social convinced me. It's the Orf7-8 deletion. This is GW.5.1.1 which had a similar deletion compared to HV.1.1, which circulated at the same time.

Comparison of age distribution of HV.1.1 and GW.5.1.1 from the US.

No, you are right. I restricted it to the US sequences to make it more apples-to-apples. Pretty clear difference.

Age distribution of HV.1.1 and GW.5.1.1

I see what you are saying, but it's not as distinct. I used HV.1.1 for comparison.

My hypothesis was incorrect. The age distributions were essentially identical.

With both viruses the fraction from kids under 10 was about 5%.

Contrast that with the US BA.3.2 sequences currently which are about 50% kids under 10.

This shift really is striking.
5/5

This image shows the age distribution of BA.2.86/JN.1 and XBBs from November 2023-January 2024. The distribution is nearly identical.

So I tested the hypothesis. I looked at the age distribution of XBB sequences vs BA.2.86/JN.1 in a subset of US samples during the 4 months when XBB was displaced.

Here are the results.
4/

However, this isn't the first time there has been a new lineage like this. The same thing happened 2 years ago when the XBBs were displaced by BA.2.86/JN.1.

If this age shift with a new lineage is 'a thing', the same thing should have then.
3/

Adults, on the other hand, have been exposed to many lineages and have broader immunity, so the difference in immunity against JN.1 vs BA.3.2 will be less.

Put together, this would mean that BA.3.2 would have a greater chance of infecting kids than adults.
2/

I had a hypothesis, and decided to test it.

Here's the hypothesis.

Kids just a few years old have probably been exposed to COVID, but the only lineage circulating has been JN.1-derivatives, so they are going to be more resistant to JN.1 than anything else.
1/

I haven't been following that closely, what's the data on stability from? Do they know what mutations drive it?

OK, I did that and tried to repost the image, and it just says that the image is missing alt text and wouldn't let me post it. I'm not sure what I'm supposed to do.

But you agree that something is different/weird. That's really my whole point.