Twenty — minutes — maybe — more.choose four words

pi day: fun

Bioinformatics and Genome Analysis Course. Izmir International Biomedicine and Genome Institute, Izmir, Turkey. May 2–14, 2016

visualization + design

Typography geek? If you like the geometry and mathematics of these posters, you may enjoy something more lettered. Visions of type: Type Peep Show: The Private Curves of Letters posters.

pi Day 2014 Art Posters

Support Ellie Balk's Kickstarter community math mural project in which Brooklyn students learn math and art to visualize pi.
2013 pi day
2014 pi day
2015 pi day
2014 pi approx day
Circular pi art

On March 14th celebrate Pi Day. Hug \pi—find a way to do it. For those who favour \tau=2\pi will have to postpone celebrations until July 26th. Some of these folks will argue that pi is wrong. If you're not into details, you may opt to party on July 22nd, which is pi approximation day (\pi ≈ 22/7).

All art posters are available for purchase.
I take custom requests.

For the 2014 pi day, two styles of posters are available: folded paths and frequency circles.

The folded paths show pi on a path that maximizes adjacent prime digits and were created using a protein-folding algorithm. The frequency circles colourfully depict the ratio of digits in groupings of 3 or 6.

Pi Day 2014 poster | Frequency distribution of digits in Pi for each of 128 6-digit groupings in 10 columns up to the Feynman Point. For each grouping the number of times a digit was seen is proportional to the width of the annulus. (zoom, BUY ARTWORK)

Pi Day 2014 poster | Frequency distribution of digits in Pi for each of 128 3-digit groupings in 12 columns up to the Feynman Point. For each grouping the number of times a digit was seen is proportional to the width of the annulus. (zoom, BUY ARTWORK)

Pi Day 2014 poster | Frequency distribution of digits in Pi for each of 128 3-digit groupings in 16 columns up to the Feynman Point. For each grouping the number of times a digit was seen is proportional to the width of the annulus. This is a very satisfying square layout. (zoom, BUY ARTWORK)

Pi Day 2014 poster | Frequency distribution of digits in Pi for each of 128 3-digit groupings in 16 columns up to the Feynman Point, with the first digit (3) offset to the top left. For each grouping the number of times a digit was seen is proportional to the width of the annulus. This is a very satisfying square layout. (zoom, BUY ARTWORK)

Pi Day 2014 poster | Frequency distribution of digits in Pi for the first 4,988 digits of Pi in groupings of 4. This subset contains the triplets for each digit, the last being 888 at digit 4,985. The layout is 29 columns and 43 rows. The first digit (3) offset to the top left. For each grouping the number of times a digit was seen is proportional to the width of the annulus. The Feynman Point 4(999999)8 is found in the middle of row 7. (zoom, BUY ARTWORK)

Pi Day 2014 poster | Frequency distribution of digits in Pi for the first 4,988 digits of Pi in groupings of 4. This subset contains the triplets for each digit, the last being 888 at digit 4,985. The layout is on an Archimedean spiral, with the the first digit (3) in the center. For each grouping the number of times a digit was seen is proportional to the width of the annulus. (zoom, BUY ARTWORK)

Pi Day 2014 poster | Frequency distribution of digits in Pi for the first 4,988 digits of Pi in groupings of 4. This subset contains the triplets for each digit, the last being 888 at digit 4,985. The layout is on an Archimedean spiral. For each grouping the number of times a digit was seen is proportional to the width of the annulus. (zoom, BUY ARTWORK)

Bayesian networks

Sun 30-08-2015

This month we continue with the theme of Bayesian statistics and look at Bayesian networks, which combine network analysis with Bayesian statistics.

In a Bayesian network, nodes represent entities, such as genes, and the influence that one gene has over another is represented by a edge and probability table (or function). Bayes' Theorem is used to calculate the probability of a state for any entity.

Nature Methods Points of Significance column: Bayesian networks. (read)

In our previous columns about Bayesian statistics, we saw how new information (likelihood) can be incorporated into the probability model (prior) to update our belief of the state of the system (posterior). In the context of a Bayesian network, relationships called conditional dependencies can arise between nodes when information is added to the network. Using a small gene regulation network we show how these dependencies may connect nodes along different paths.

Puga, J.L, Krzywinski, M. & Altman, N. (2015) Points of Significance: Bayesian Statistics Nature Methods 12:277-278.

Puga, J.L, Krzywinski, M. & Altman, N. (2015) Points of Significance: Bayes' Theorem Nature Methods 12:277-278.

Unentangling complex plots

Fri 10-07-2015

The Points of Significance column is on vacation this month.

Meanwhile, we're showing you how to manage small multiple plots in the Points of View column Unentangling Complex Plots.

Data in small multiples can vary in range, noise level and trend. Gregor McInerny and myself show you how you can deal with this by cropped and scaling the multiples to a different range to emphasize relative changes while preserving the context of the full data range to show absolute changes.

McInerny, G. & Krzywinski, M. (2015) Points of View: Unentangling complex plots. Nature Methods 12:591.

Fixing Jurassic World science visualizations

Fri 10-07-2015

The Jurassic World Creation Lab webpage shows you how one might create a dinosaur from a sample of DNA. First extract, sequence, assemble and fill in the gaps in the DNA and then incubate in an egg and wait.

We can't get dinosaur genomics right, but we can get it less wrong. (a) Corn genome used in Jurassic World Creation Lab website. Image is from the Science publication B73 Maize Genome: Complexity, Diversity, and Dynamics. Photo and composite by Universal Studios and Amblin Entertainment. (b) Random data on 8 chromosomes from chicken genome resized to triceratops genome size (3.2 Gb). Image by Martin Krzywinski. (c) Actual genome data for lizard genome, UCSC anoCar2.0, May 2010. Image by Martin Krzywinski. Triceratops outline in (b,c) from wikipedia.

With enough time, you'll grow your own brand new dinosaur. Or a stalk of corn ... with more teeth.

What went wrong? Let me explain.

Corn World: Teeth on the Cob.

Printing Genomes

Tue 07-07-2015

You've seen bound volumes of printouts of the human reference genome. But what if at the Genome Sciences Center we wanted to print everything we sequence today?

Curiously, printing is 44 times as expensive as sequencing. (details)

Gene Volume Control

Thu 11-06-2015

I was commissioned by Scientific American to create an information graphic based on Figure 9 in the landmark Nature Integrative analysis of 111 reference human epigenomes paper.

The original figure details the relationships between more than 100 sequenced epigenomes and genetic traits, including disease like Crohn's and Alzheimer's. These relationships were shown as a heatmap in which the epigenome-trait cell depicted the P value associated with tissue-specific H3K4me1 epigenetic modification in regions of the genome associated with the trait.

Figure 9 from Integrative analysis of 111 reference human epigenomes (Nature (2015) 518 317–330). (details)

As much as I distrust network diagrams, in this case this was the right way to show the data. The network was meticulously laid out by hand to draw attention to the layered groups of diseases of traits.

Network diagram redesign of the heatmap for a select set of traits. Only relationships with –log P > 3.9 are displayed. Appears on Graphic Science page in June 2015 issue of Scientific American. (details)

This was my second information graphic for the Graphic Science page. Last year, I illustrated the extent of differences in the gene sequence of humans, Denisovans, chimps and gorillas.

Sampling distributions and the bootstrap

Thu 11-06-2015

The bootstrap is a computational method that simulates new sample from observed data. These simulated samples can be used to determine how estimates from replicate experiments might be distributed and answer questions about precision and bias.

Nature Methods Points of Significance column: Sampling distributions and the bootstrap. (read)

We discuss both parametric and non-parametric bootstrap. In the former, observed data are fit to a model and then new samples are drawn using the model. In the latter, no model assumption is made and simulated samples are drawn with replacement from the observed data.

Kulesa, A., Krzywinski, M., Blainey, P. & Altman, N (2015) Points of Significance: Sampling distributions and the bootstrap Nature Methods 12:477-478.