iiinova — where your digital self keeps evolving

Kamil Slowikowski @slowkow.com · 15/08/2026

Maryland Crashes from 2018-2026 Data from mdsp.maryland.gov run.cosmograph.app/public/e09a3...

Visualization of Maryland State Police crash data from 2018–2026 on a dark background. Thousands of color-coded points form dense, interconnected road networks across Maryland, with the greatest concentration in the Baltimore–Washington corridor. Labels identify prominent crash locations and roads.

000

Kamil Slowikowski @slowkow.com · 01/12/2025

Cruz Godar shares stunning Javascript and webGL applets that demonstrate the beauty of mathematics. cruzgodar.com

020

Kamil Slowikowski @slowkow.com · 02/10/2025

A visualization of average resting heart rate in Germany from 2020 through 2022. Link: corona-datenspende.github.io/en/vitaldata... Data: zenodo.org/records/8229...

020

Kamil Slowikowski @slowkow.com · 23/09/2025

While many immunologic abnormalities in acute severe COVID-19 resolve during convalescence 3-months post-infection, we observed persistently high ICOS expression in regulatory T cells, potentially linking acute infection to chronic post-COVID syndromes.

**M.** Volcano plot for convalescent CD4 T cell subset 6 (Treg).
Dots indicate genes, x-axis indicates fold-change, y-axis indicates negative log10 p-value from multivariate linear regression.
Red color indicates FDR < 5%.

**N.** Box plot of *ICOS* gene expression in CD4 T cell subset 6 (Treg) in the convalescent data.
Fold-change and p-value is shown.
Dots represent patients, box plots indicate median and interquartile range, x-axis and hue represent convalescent COVID infection.

**O.** Box plot of *ICOS* gene expression in CD4 T cell subset 13 (Treg) in the acute data.
Dots represent patients, box plots indicate median and interquartile range, x-axis represents time point, and hue represents COVID infection.

121

Kamil Slowikowski @slowkow.com · 23/09/2025

We also identify HLA-DQB1 amino acid positions associated with: - COVID-19 disease severity - specific TCRs in CD8 T cells - viral load - neutralization capacity of serum

**E.** Amino acid position associations with COVID severity for HLA-DQB1.
x-axis indicates position along the gene.
y-axis indicates negative log10 p-value from multivariate linear regression.
Black color indicates FDR < 5%.

**F.** Coefficients from the multivariate linear regression for the effect size of each amino acid position on COVID severity.
Amino acid positions D57 and V57 are highlighted.
Error bars indicate 95% CI.

**G.** Bar plots of the number of patients with each genotype.
Facets indicate genotype (0, 1, 2) for amino acid position D57 (top) and V57 (bottom).
x-axis indicates number of patients.
Hue indicates COVID severity.

**H.** Box plot of abundance of CD8 T cells with TRBV28, x-axis indicates time point and hue indicates genotype of HLA-DQB1 V57 (0, 1, 2).
P-value from multivariate linear regression.

131

Kamil Slowikowski @slowkow.com · 23/09/2025

We found that tocilizumab eliminates CLU-expressing MDSCs and ISG-positive myeloid subsets, restores antigen presentation, and reactivates productive adaptive immunity. In myeloid cells, the mRNA signature of tocilizumab treatment appears inverse to the signature of disease severity.

A scatter plot where each point represents a gene. The x-axis represents the fold-change of the gene in acute patients who received tocilizumab treatment, where genes induced after treatment are seen on the right and genes repressed after treatment are seen on the left. The y-axis represents the fold-change of the gene in acute patients, where genes associated with worse severity are seen on the top, and genes associated with lesser severity are seeon on the bottom. Some genes are labeled.

131

Kamil Slowikowski @slowkow.com · 23/09/2025

Severe disease is also linked to autoantibodies targeting type I interferons, influenced by specific HLA-DQB1 allelic variants, and strongly correlated with serum IL-6 levels. Here is a schematic representation of our findings:

A schematic that shows some factors associated with the severity of COVID-19. Starting with SARS-CoV-2 infection, the adaptive immune response should lead to successful elimination and convalescence. In some patients, various factors such as viral titer, age, autoantibodies, lymphopenia, serum IL-6 levels, TCR and BCR clones, and HLA genotypes can increase risk for severe disease. This leads to a feedback loop wherein lack of viral clearance leads to increased tissue damage. This is associated with myeloid cell dysfunction marked by impaired antigen presentation, which drives a non-productive adaptive immune response, as reflected by reduced expression of B and T cell gene programs involved in antigen recognition, immune synapse formation, and cytotoxicity. The anti-IL-6R antibody, tocilizumab, appears to reverse the mRNA signature of disease severity in myeloid cells.

133

Kamil Slowikowski @slowkow.com · 23/09/2025

I'd like to announce the medRxiv preprint of our latest work: A multimodal atlas of COVID-19 severity identifies hallmarks of dysregulated immunity. www.medrxiv.org/content/10.1...

Overview of a research study about COVID-19. The study includes three cohorts: (1) 351 acute COVID-19 patients across 5 severity levels and 3 time points profiled by multiple modalities, (2) 4 acute COVID-19 patients treated with tocilizumab, (3) 73 patients, 43 of whom were previously infected with SARS-CoV-2 3 months ago. The overview shows the general study design and visualizations of the single-cell RNA sequencing data analyzed in this study.

1103

Kamil Slowikowski @slowkow.com · 13/09/2025

Search #Pubmed and view the results on a 2D embedding. Each point represents a paper, and similar papers are near each other. Colors indicate clusters of similar papers. By Viktor Petukhov github.com/multicore-ca...

040

Kamil Slowikowski @slowkow.com · 08/08/2025

1. Make an Excel file with your author information 2. Drag and drop 3. Copy a nicely formatted author list into Word 🐶 slowkow.com/authorbud

031

Kamil Slowikowski @slowkow.com · 05/08/2025

"After adjusting for smoking and other risk factors, we observed statistically significant and robust associations between air pollution and mortality." "Air pollution was positively associated with death from lung cancer and cardiopulmonary disease" pubmed.ncbi.nlm.nih.gov/8179653/

A graph from a 1993 article published in the New England Journal of Medicine illustrating the relationship between air pollution and mortality rates in six U.S. cities. The x-axis represents the concentration of fine particles (µg/m³), while the y-axis shows the adjusted mortality-rate ratios. Cities such as Steubenville, OH, and Portage, WI, are marked. The adjusted mortality rate for the most polluted of the cities (Steubenville, OH) is 1.26 times higher than the morality rate least polluted city (Portage, WI). Data from Dockery et al., NEJM 1993.

021

Kamil Slowikowski @slowkow.com · 16/07/2025

Convert Markdown to C++, Python, MATLAB, and LaTeX iheartla.github.io by Yong Li

050

Kamil Slowikowski @slowkow.com · 06/03/2025

Airborne Infection Prevention Using Germicidal UV-C: Training and Certification Program August 2025 in College Park, Maryland Developed by Dr. Rolf Bergman, Dr. Paul Jensen, Dr. David Sliney Funded by Balvi Filantropic Fund (@vitalik.ca) Sign up at: docs.google.com/forms/d/e/1F...

044

Kamil Slowikowski @slowkow.com · 20/12/2024

"December 16, 2024: Transmission is higher than during 73% of the pandemic, and lower than during 27% of the pandemic. We are seeing about 5 million infections a week now, likely resulting in over 200,000 eventual Covid-associated conditions or Long Covid." www.pmc19.com/data/index.php

US New Daily COVID-19 Infections and 1-Month Forecast

"As of December 16, 2024: 1.6% of the population (1 in 64 people) is estimated to be actively infectious with Covid, meaning just under 750,000 daily infections. Transmission is higher than during 73% of the pandemic, and lower than during 27% of the pandemic. We are seeing about 5 million infections a week now, likely resulting in over 200,000 eventual Covid-associated conditions or Long Covid."

Source: PMC19.com/data by Michael Hoerger, PhD, MSCR, MBA, @michael_hoerger

Year-Over-Year Comparison of US COVID-19 Transmission. Source: PMC19.com/data

"2024 transmission has varied considerably from prior years. Notice the large, late, and atypically shaped summer wave, the extremely rapid
decline in transmission post-wave, the extended lull, the lack of increase in transmission in November, and now the extremely large increase in transmission in December.

The level and timing of the wave peak remain highly uncertain, as does the level of transmission post-peak. If transmission continues to pick up at rates much higher than average, we could still see a peak around New Year’s Eve – a “silent surge” in which we got from lull-to-peak faster than ever before, many are caught off guard, and are shocked to get Covid during holiday travel, family gatherings, and back-to-school. Alternatively, a peak closer to January 7 remains plausible. We are in uncharted territory in terms of the pattern of transmission and behaviorally testing how quickly as a society we can increase transmission."

2386

Kamil Slowikowski @slowkow.com · 16/12/2024

#ggrepel lives in RWorld, which shares a border with Coronaland, Celltopia, and Genomic Haven. #maplibre #github

150

Kamil Slowikowski @slowkow.com · 03/01/2024

COVID screening of pooled samples from 7 international airports from 6 US States: · California, New York, Washington, New Jersey, Georgia, Virginia Positivity rate: 23.5% (116 / 494 pools) on Dec 10, 2023 Source: covid.cdc.gov/covid-data-t... Chart: observablehq.com/@slowkow/cdc...

A chart showing the COVID positivity rate for samples from 7 international airports from 6 US States. Each line represents data from one year, including 2021, 2022, and 2023. The latest data shows that 23.5% of pools (representing 6,086 total participants) were positive for COVID. The data source is CDC Genomic Surveillance, available at the following URL: https://covid.cdc.gov/covid-data-tracker/#traveler-genomic-surveillance

020

Kamil Slowikowski @slowkow.com · 18/09/2023

Sometimes I want to know how many PubMed results I will get for all pairs of search terms. Maybe I'm researching viruses, so I want to run each one of these queries: · virus IL6 · virus IL8 · virus IL10 · and so on... I made a simple web app for this: slowkow.github.io/pubmed-pairs/

A screenshot of a website called PubMed Pairs. There are two input text boxes, one for each list of search terms. The app automatically searches for each pair of terms and shows the number of PubMed results for each pair. Titles and abstracts are shown below.

030