iiinova — where your digital self keeps evolving

Conor

@lwlss.net

Former research scientist, interested in intersections between maths, computing, biology, engineering, art and music | A' fuireach ann an Uibhist a Tuath agus ag ionnsachadh na Gàidhlig | cnr.lwlss.net

Posts Replies Media

Conor @lwlss.net · 29/05/2025

000

Conor @lwlss.net · 29/05/2025

000

Conor @lwlss.net · 16/05/2025

022

Conor @lwlss.net · 20/02/2025

Naive comparisons between cells from healthy subjects and patients leads to very large overestimates of the proportions of affected cells, which leads to poor measurements of disease severity.

Scatterplots for visualising protein abundance per cell in order to classify cell OXPHOS status. x-axis values represent the abundance of general mitochondrial proteins, whereas y-axis values represent the abundances of specific proteins which we believe are affected by mitochondrial diseases. Points labelled yellow represent cells that have been manually annoted by experts as defective. Red points represent cells that have been classified as defective by a naive, historical approach. Blue points represent cells that have been classified as normal by the same approach.

000

Conor @lwlss.net · 20/02/2025

New paper out today in @PLOS computaional biology Bayesian classification of OXPHOS deficient skeletal myofibres The progression of many types of mitochondrial disease occurs randomly in each of the (trillions of) cells that make up the patient. doi.org/10.1371/jour...

Microscopy image of patient muscle fibre cross sections stained to show the abundance of a mitochondrial protein. The mosaic pattern of abundance is a result of the random nature of mitochondrial disease progression in individual patient cells.

110

Conor @lwlss.net · 14/01/2025

Some yellow 1D videos on vimeo. vimeo.com/1046801427 vimeo.com/1046801345

Thousands of very thin horizontal yellow stripes. Colours are slightly different to each other and adjacent colours are related.

Thousands of very thin vertical yellow stripes. Colours are slightly different to each other and adjacent colours are related.

000

Conor @lwlss.net · 11/01/2025

000

Conor @lwlss.net · 11/01/2025

4K versions of these videos available from vimeo: vimeo.com/1045788310 vimeo.com/1045788384

100

Conor @lwlss.net · 10/01/2025

Some more image generation using Julia. Revisiting 1D videos. I've written about them before: cnr.lwlss.net/1DMoovies/ cnrlwlss.github.io/1D/ Large collection of them on vimeo: vimeo.com/showcase/159... The latest ones are in 4K (see link above). I'll add some HD versions frames here too.

Vertical stripes, gradually changing colours.

Horizontal stripes, gradually changing colours.

100

Conor @lwlss.net · 07/01/2025

000

Conor @lwlss.net · 03/01/2025

Julia also plays well with the very powerful ffmpeg library for video editing, so you can build a video from adjacent sample images such as the one above. The VideoIO.jl library allows for iterative construction of video, which saves a lot of memory! Output attached. gist.github.com/CnrLwlss/78a...

110

Conor @lwlss.net · 03/01/2025

It's also nice to work with existing images in Julia. The snippet of code at the link below downloads an image, reads it to memory, cuts an 8x8 square of colour pixels out of the image, expands the image as above and saves to file. Output attached. gist.github.com/CnrLwlss/5a1...