Twitter | Search | |
Claire McWhite
Grad student at UT Austin. Prev Rice. Protein complexes, Systems Bio, comparing things to each other. Status: Cilia, plants. Marcotte lab.
1,480
Tweets
524
Following
365
Followers
Tweets
Claire McWhite 9h
The fabled cajun turducken.
Reply Retweet Like
Claire McWhite retweeted
Gabriel N. U. Nov 22
If that awful baby “Velociraptor” from the new left a bad taste in your mouth, here is what likely looked like when alive. You can get the poster I did showing Velociraptor and his close relatives at
Reply Retweet Like
Claire McWhite retweeted
Allister Crow Nov 21
The MacB crystal structure has just gone full augmented reality! Check it out for yourself, by pointing the 'Augment' app at the front page of the Koronakis lab's PNAS paper!
Reply Retweet Like
Claire McWhite retweeted
Chlamy Autophagy Nov 20
Amazing microscopy study showing the structure of COPI coat within Chlamydomonas cell
Reply Retweet Like
Claire McWhite Nov 19
Guilty After 3 years now at a cow-themed university, I probably should have picked up some general bovine knowledge.
Reply Retweet Like
Claire McWhite retweeted
Matthew Dalby, PhD Nov 18
I love these character studies of Neanderthals by the artist Tom Björklund. So rarely are they pictured as if they were real people. See his blog and Facebook for more.
Reply Retweet Like
Claire McWhite Nov 19
Replying to @ClausWilke
Real. It's the 'Cows' dataset from the 'PASWR' package. I found it here:
Reply Retweet Like
Claire McWhite Nov 19
Reply Retweet Like
Claire McWhite retweeted
Pastel BioScience Nov 19
1809 | Asparagine isolated and named, due to ready crystallization from asparagus shoots. Wasn't recognized as an amino acid till 1873 |
Reply Retweet Like
Claire McWhite Nov 17
Replying to @jonmlaurent
See #17, the newly discovered void in the great pyramid. Yes new journal clubs
Reply Retweet Like
Claire McWhite Nov 17
Replying to @wytham88
Also some illuminati stuff.
Reply Retweet Like
Claire McWhite Nov 17
Replying to @wytham88
Each person in the room briefly describes 1-2 papers. Someone summarizes on the board. Then, we take a 1st vote for for what we want to hear more about (unlimited votes/person). The top few papers go the 2nd round (1 vote/person). Then someone presents last weeks' winning paper
Reply Retweet Like
Claire McWhite Nov 17
Marcotte lab journal club! Winner by far was the fungus that bypassed ant brains. My favorite was an improved nearest neighbor algorithm learned from c elegans odor sensing
Reply Retweet Like
Claire McWhite Nov 17
I don't know what this experiment is, but it looks like an awesome setup!
Reply Retweet Like
Claire McWhite Nov 16
I guess I'll just parse it. Though I do now have a segmentation for bitmaps. readImage("spine.png") %>% bwlabel() %>% matrix(nrow=1508, ncol=2794, byrow=F) %>% melt %>% filter(value>1) %>% ggplot(aes(x=X1, y=X2, color = as.character(value), group=value)) + geom_point()
Reply Retweet Like
Claire McWhite Nov 16
I wanted to avoid xml parsing...but that is a way. Just a test image. I also just realized that the raster conversion is totally unnecessary, I just need to reshape the matrix and filter.
Reply Retweet Like
Claire McWhite Nov 16
I just need to stop raster() from filling in the outlines, then I can use geom_polygon with fill. I'm using the EBImage and raster libraries btw. Filling in colors for values manually in inkscape pains me conceptually
Reply Retweet Like
Claire McWhite Nov 16
Alright, almost there! readImage("spine.png") %>% bwlabel() %>% as.array() %>% matrix(nrow=2794, ncol=1508, byrow=T) %>% raster() %>% rasterToPoints(function(x) x %in% seq(2,28)) %>% data.frame %>% ggplot(aes(x, y, color=as.character(layer), group=layer)) + geom_point()
Reply Retweet Like
Claire McWhite Nov 16
Goal: Start with a picture of a spine for example, and color each vertebra by values. Like mapping colors onto states, but with a different image. I don't know atm how to get the groups to color vertebra by with just the inkscape coordinates.
Reply Retweet Like
Claire McWhite Nov 16
I haven't tried any of this but consider a bw image of a spine... 1. bwlabel(img) gives pixels in each vertebra a value 2. Extract each pixel value into separate images 3. imagemagick edge detect gives coords for each vertebra 4. Recombine and plot with geom_polygon.
Reply Retweet Like