Chenxin Li, PhD (@chenxinli2.bsky.social)
@ChenxinLi2
Followers
9,901
Following
2,923
Media
1,440
Statuses
11,507
Li | Assistant Research Scientist @ UGA | Decoding how genome, development, and environment shape (specialized) metabolism | Also post data visualization
Joined July 2019
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
México
• 425541 Tweets
#BETAwards
• 409392 Tweets
#HouseOfTheDragon
• 210651 Tweets
Ecuador
• 182331 Tweets
Usher
• 112091 Tweets
Rhaenyra
• 110407 Tweets
#ForbiddenDoor
• 98092 Tweets
EMMANUEL
• 90391 Tweets
Alicent
• 80850 Tweets
CHANGO MAS FURIOSO
• 73126 Tweets
Coco
• 73024 Tweets
Delgado
• 72501 Tweets
Vinotinto
• 71371 Tweets
Clippers
• 65635 Tweets
Bautista
• 57473 Tweets
Aegon
• 46068 Tweets
Chris Paul
• 44524 Tweets
Taraji
• 41473 Tweets
#HOTD
• 41204 Tweets
Tyla
• 35298 Tweets
Chloe
• 32520 Tweets
Chris Brown
• 31628 Tweets
Aemond
• 25970 Tweets
Teyana
• 25221 Tweets
Teyana
• 25221 Tweets
Will Smith
• 23637 Tweets
Keke
• 21421 Tweets
幻想シアター
• 17669 Tweets
Concacaf
• 17383 Tweets
Wemby
• 17279 Tweets
Viserys
• 17097 Tweets
Mechico
• 15731 Tweets
Childish Gambino
• 13759 Tweets
だいち4号
• 13298 Tweets
Jimmy Lozano
• 13078 Tweets
桃の天然水
• 12111 Tweets
Lauryn Hill
• 11604 Tweets
打ち上げ成功
• 11082 Tweets
Naji
• 10668 Tweets
Marsha
• 10517 Tweets
Antuna
• 10472 Tweets
Tinashe
• 10087 Tweets
Pinned Tweet
Look what came in the mail today
19
2
366
Weird microbe on a pasta sauce jar lid! My student returned from her trip and found this purple pink ish flower looking thing in her fridge. Any idea what this might be?
@ContamClub
312
453
6K
Not enough people know about
Open source vector graphics. Free alternatives to BioRender.
6
564
2K
I am not sure. If I were a PI I would try to preserve my mentee’s writing as much as possible-light edits only. If substantial changes are needed, I leave a comment instead. This is their writing, and they should express their individuality, not mine.
If I could give aspiring scientists and trainees one piece of advice it would be this: Find a mentor who cares enough about you to do this 👇
#Mentorship
#AcademicChatter
#PhysicianScientist
#CareerAdvice
212
730
6K
112
111
2K
Returned to the lab bench and found out there is mold? growing in my 10% Triton. Dude, you are living in literal detergent.
81
74
2K
Yesterday we had a Zoom lunch w/ a former PhD student of the lab. He said he has been using my GitHub repo to brush up on his R skills. "It's well put together" he said. I'm so glad to hear that! I wrote these when I was slacking off in the lab. 😂
8
271
1K
Today I finished teaching "quick data vis" to 1st year grad students the 2nd time! Lessons/scripts + example data available:
Online R learning repository main page (basic statistics focus):
Quick data vis page (data vis focus):
10
286
1K
Is it a general phenomenon that PhD students become petty, salty, and jaded by year 5-6? Asking for a friend.
147
43
1K
Update: The package has been dropped off at UPS. I used a spatula to transfer the purple stuff to a 50 mL conical tube (+ a bit of sauce to keep it alive), packed in a styrofoam box w/ 2 ice packs, then into a cardboard box. Should be delivered to Dr.
@ImperfectFunGuy
Monday!
22
21
1K
Super nice when people asked me if they can use my repository for teaching. I always say yes.
Friends don’t let friends make bad graphs repository:
14
151
1K
I remember someone from a highly accomplished lab told me they would leave out 1-2 key pieces of data/experiment for their 1st manuscript draft, so that reviewers have low hanging fruit to pick instead of trying to nitpick. While they wait for reviews, they do those experiments.
37
73
1K
I learned how to use ggtreeExtra this week for phylogeny visualization and annotation.
ggtree:
ggtreeExtra:
8
198
1K
Garlic genome is almost 17Gb?! Oh wow.
I was prepping dinner and chopping some garlic and went like “I wonder what’s the genome size?”
24
127
1K
You can make an entire figure in R using the patchwork package. A tutorial is available at my GitHub to generate this exact figure (). It covers controlling layout, relative heights/widths of panels, hierarchical assembly, and position of letter labels.
17
205
1K
Does grad school make people depressed or do depressed people tend to apply for grad school?
25
50
1K
A new tutorial on how to reorder rows & columns of a heatmap has been added to my "Friends don't let friends make bad graphs" repository.
Link to tutorial:
Link to "Friends don't let friends make bad graphs":
3
193
1K
What to use my scripts for gene co-expression analysis or generate nice graphics, but no experience with R? Don't worry, I have you covered! I have separate "Online_R_learning" tutorial on GitHub. It has 14 lessons, but really just the first 5 will get you a long way.
11
216
1K
The wet lab scientists think I am a bioinformatician. The bioinformaticians think I am a wet lab scientist. But I am fundamentally just a grapher. People hand me data and I make pretty graphs for them and help them interpret their results.
21
39
1K
This Fall I'll be teaching "quick data visualization" as part of a 1st year grad student seminar again for PBGG at UGA. The graphical abstract has 7 sections, but an 8th section on figure assembly has been added as well.
Contents available on GitHub:
14
200
1K
Telomere
18
68
889
Sloppy is good. If you get reliable results with sloppy execution, it means your method is robust and tolerant to errors. Precision is overrated.
15
77
798
If you make all the panels in R, you can actually make the entire figure in R with the patchwork package ().
11
118
758
Today I start teaching the "quick data vis" module to 1st year graduate students. This is my 2nd time teaching it.
GitHub:
6
129
758
"We are looking for a motivated individual to fill this postdoc position." Actually, nowadays if someone wants to be a postdoc, they are already very motivated, okay?
16
45
746
Gonna try out this gganimate package ()
Not really compatible with publications but would be great for presentations. Reminds me of Hans Rosling's presentations.
9
69
738
I made a new tutorial: "Simple Tidy GeneCoEx", a simple gene co-expression analyses workflow powered by tidyverse and graph analyses. Link:
Also graphics generated from an example dataset:
15
159
736
Just found out there has been massive update(s) at Bioicons! () Free vector graphics for biology, now include extension to Inkscape and online drawing user interface. Highly recommend!
7
182
709
Shameless plug: I wrote a R tutorial (for my undergrad interns at the time). It covers basic tidyverse (data wrangling & visualization), + basic applied stats (all sort of ANOVA/regression). Math kept to minimum, focuses on interpretation & execution in R.
4
158
674
NIH proposes upping postdoc salary to 70k/year.
People: This will lead to less postdocs!
Me: Didn't you see how many labs are struggling to hire a postdoc? Very few PhD students are even looking to be a postdoc. It's gonna be okay. Chill.
25
37
624
Heatmap can be the worst data visualization if you aren't doing it right. For example, reordering/grouping rows & columns gives a completely different look to a heatmap from the same data.
Tutorial for programmatically re-order rows/columns here:
5
77
574
Beware of bimodal distributions in box plots, for example group B in this example. I would suggest instead of box plots and violin plots, use bee swarm plots (shown on the right).
💡A boxplot can hide information. Show it with a violin plot instead?
(Trying to visually explain this 😛)
46
551
4K
10
100
599
Finished teaching my 1st course! I was co-instructor for 1st year grad student seminar "Grad School Life Skills", where I was allotted 7 weeks to cover data visualization. Course material are available online under my "Online_R_learning" repo:
3
114
593
Here's my version of "Friends don't let friends make bar plots" for showing means separation of replicated experiments.
9
190
576
Friends don't let friends forget to reorder stacked bar plots. Stacked bar plots are commonly used in research (e.g., to visualize admixture analysis). It can get messy! After reordering the bars, wow, it really made a difference, don't you think?
Link:
5
80
573
Hello new followers! You are probably here for Purple Pasta Sauce. I'm a plant biologist studying medicinal plants. I look at their basic biology (genetic makeup, metabolism, development) and create biotech tools for them. Come for Purple Pasta Sauce and stay for plant science!
18
24
572
Another rendition on "friends don't let friends make bar plot", or at least check a couple things before settling down on bar plots.
In this example, two groups have similar means and standard deviations, but quite different distributions. Are they really "the same"?
7
98
566
Our plant nuclei isolation protocol is online! Frantically chopping on ice releases nuclei from cells, and a key factor for nuclei integrity is the detergent concentration. We tested & optimized the protocol for a range of species, including major crops.
17
127
530
Here is "friends don't let friends make heatmap without checking outliers". This applies to using colors to represent numeric data in general. Having outliers in heatmap can really change how we perceive & interpret the visualization.
6
60
522
Don’t join a lab that emphasizes “passion” or “motivation” because they are likely looking to make you work more hours. Instead, join a lab that emphasizes planning, prioritizing, and work-life balance.
As a trainee, you may often feel pressure to conform to this "passion"-driven lifestyle of putting work over health, relationships, and hobbies. I 100% respected my friend's hustle but it's also okay if you're like me - you can still do well and end up where you want to be!
3
35
691
11
71
518
OMG… Now I got stuff growing in my 10% Tween20. What’s with microbes growing in literal detergent? Should I try to isolate it (spread it on a plate with just agar, 10% Tween20 and a pinch of yeast extract)?
69
11
506
Me: talk about existential dread being a postdoc.
People: it doesn't end after you become a PI.
Me: Thanks. But think about it this way. Postdocs want to be PI, but no PI wants to come back and be postdoc again. Obvious one is better than the other. LMAO. 🤷♀️
39
19
497
My "Friends don't let friends make bad graphs" repository has reached 1000+ stars on GitHub! I can't believe it! Thanks everyone for the interest and support. Gonna add this to my CV. 😂😂😂
8
76
490
First author here. I have a step-by-step tutorial for this Tidyverse-based gene coexpression workflow on GitHub:
Discovering new insights in genetics! 🧬 "Simple Tidy GeneCoEx" is a customizable gene co-expression analysis tool that uses R programming and igraph functions to detect clusters of genes with similar expression patterns.
Read
#OA
in
@plantgenome
👉
1
17
73
10
95
477
My Online_R_learning tutorial got to 200 stars today! (It includes the "quick data vis" sub-module, which I think most of the new stars are for.)
2
74
475
Same gradient setup, different species. The density of chloroplast and nuclei are different across species.
8
27
469
Small milestone: Online_R_learning has reached 400 stars on GitHub!
My repositories have a much farther reach than my actual research:
Friends don't let friends make bad graphs: 3.8k stars
SimpleTidy GeneCoEx: 154 stars.
5
65
452
Preprint of our tidyverse+graph theory based gene coexpression analysis workflow is online! Tidyverse functions produce nodes & edges, and graph based clustering (Leiden) detects tight gene coexpression modules.
Tutorial:
4
105
445
People warned me not to share ideas too openly on Twitter/at conferences, because I might get scooped. They are probably right, but I also specialize in high-risk low-reward project ideas, so really no sane person would take my ideas.
9
8
426
Visiting Texas A&M for an on-site interview in the biochem department. Wish me luck!!!
34
2
427
What are your lab culture shock experience? In my PhD lab, trainee(s) write the entirely of the paper, and the PI approves it. I.e., trainees generates >99% of the text. I was quite shocked to know that in many labs the PI generates >99% of the text in publications.
37
14
423
Some PIs are rude. I had this experience multiple times at conferences. I was chatting with X, and prof Y came along, interrupted our conversation, starting to talk to X as if I wasn't there. Dude, how hard is it to at least say "sorry, I need to talk to X"?
21
8
423
In case anyone is interested, I also have ggtraces, a ggplot compatible tool for visualizing line traces (e.g., chromatograms), which aligns line graphs along a parallelogram. A key feature is adjustable perspectives between side view & top view. GitHub:
6
58
383
"Friends Don't Let Friends Make Bad Graphs" repository.
5
50
372
I was mentally prepared for this 11-kb PCR to not work. But it did. From let to right temperature gradient of annealing temperature from 62 to 72.
15
14
363
Recombinant protein purification simplified to passing the culture through a 0.45 um filter. Incredible stuff.
From:
0
64
353
I have been upgraded to the rank of Assistant Research Scientist in the Department of Crop & Soil Sciences at University of Georgia. As a non-tenure-track research faculty, my new title will allow me to submit grants as the PI.
26
8
351
Graphical abstract for my new "Quick data vis" sub-repository within "Online_R_learning". Content still under construction. First 4 lessons already completed. Stay tuned.
3
74
349
Bad news: Reviewer
#2
would not accept the manuscript no matter what even though Reviewers 1&3 has recommended its publication.
Good news: Friends don't let friends make bad graphs have 500+ stars!
8
48
335
Tm calculators & optimal annealing temperatures for PCR is a myth (for the most part). Each set of 8 is a temperature gradient of annealing temp ranging from 55-70C. As you can see all tested temperatures look the same.
32
20
319
Say goodbye to VennDiagram and UpsetR...Customized upset plots are pretty and clear. Upper left: total set sizes (geom_bar); upper right: legend (get_legend() from cowplot ; lower left: colored matrix showing intersections (geom_tile); lower right: intersection sizes (geom_bar).
9
84
323
One of the biggest challenges of being a postdoc is I never know if I have done enough. 1/🧵
16
13
303
All 7 lessons of "Quick data vis" have been prepared and uploaded.
You can download the repository, follow the text in the R scripts, and run them as you read.
Or you can browse the lesson as html online in the Lessons folder on Github.
1
57
289
People say the nanodrop is a random number generator, but if you think about it...Every instrument is a random number generator, each with a different underlying distribution that we randomly sample from.
18
16
287
A stable release of ggpathway, a ggplot compatible pathway visualizer, is live on GitHub and has been pushed to Zenodo.
GitHub:
Feature requests and feedback welcome.
4
50
273
Funny situation: two plasmids from Addgene, total $170 ($85 each), not counting shipping.
Just synthesizing the pieces that I need from Twist: 400-bp piece and 1000-bp piece, add up to about $100.
I might just order synthesis.
18
7
276
My account could be suspended from posting this. Figure from:
6
40
268
Sometimes I wish more postdocs are on Twitter. Last cycle I applied to 15 TT positions and got 0 interviews. Obviously my CV was way too weak. But I am curious who got the job, because the successful candidates would have announced it on their Twitter.
21
11
267
Just talked with a good friend who is finishing up his PhD.
Me: “is your professor trying to keep you for another paper?”
Friend: “no, but I want to finish another paper or at least a preprint. Because if I leave it unfinished, he won’t bother to publish it.”
Wow. 🥲🤯😑
17
12
264
The problem w/ UMAP edgelords is the disregard for known biology. A simple exercise is take a few experimentally validated cell type specific genes (there are many), and ask how often do cells expressing the same cell type determinant 'happen to' be close together on a UMAP.
16
25
262
I think GitHub repositories should count towards scholarly products or impacts or whatnot. I am not going to try to publish my "Friends don't let friends make bad graphs" repo, but geez, it has like 220 stars as of today.
8
37
258
Plant transformation is going to be the biggest bottleneck for plant synthetic biology. All these exciting gene circuits are demonstrated in bacteria or Arabidopsis or transient expression. I’m surprised funding agencies haven’t thrown more money into plant transformation.
18
23
251
[Shit post warning]
.
.
.
.
.
.
I am quite disappointed by how people cheered Texas A&M for removing DEI statements in hiring. These people include plant scientists that I know. A thread
🧵1/n.
9
42
247
How does your lab name your freezers? We named ours after organelles.
32
11
240
Happy to announce that our paper on single cell omics (RNA + metabolites) for medicinal plant Catharanthus is officially online. This is a fruitful collaboration between
@OConnor_lab
&
@TheBuellLab
. This is my first paper for my postdoc work.
10
45
234
True wasabi plants are very rare, largely restricted to Japan. Wasabi condiment nowadays are made with horseradish and dyed green. I wonder if I can do cell culture with wasabi and make real wasabi sauce from culture. 🧐🧐🧐
15
17
232
Lab mate: "Li can you send me your protocols regarding cloning".
Me: "Eh, I do things out of muscle memory but I can type it out for you later or I can just show you how to do it."
13
13
230
Today I am officially upgraded to Dr Li. What a journey.
18
0
228
The documentations of many bioinformatics tools are so bare minimum (or even outright bad) that perhaps only the lab or person who wrote them can use the tool. Not gonna name names. But a GOOD model is bedtools, which has awesome documentations.
4
15
224
Side note: instead of lab grown meat, I would like lab grown olive oil. Olive oil straight from olive cell culture. Olive oil supply independent of climate would be nice.
10
19
221
My Friends don't let friends make bad graphs repository has reached 6000+ stars!
0
22
222
My "Friends don't let friends make bad graphs" series is somewhat complete. I think I covered the most common and egregious mistakes in graphing. Let me know if you see something egregious and I will add it to the series.
7
78
216
UC grad students requested cost of living wage adjustment. UC denied the request: "it's not through a union".
This year, a majority of UC grad students + fellows petitioned to unionize.
UC did not recognize the majority: "trainees are not employees". What a clown show. 🤡🤡🤡
8
46
212
Having fun playing w/ gganimate ().
Watch 200 years of data from the Gapminder foundation ().
x = income (log10 scale)
y = child mortality / 1000 born
color = birth / woman
1 frame = 1 year
4
14
204
Double CRISPR knockout of pectin degrading enzymes improves tomato shelf-life while ensuring fruit quality
Really nice figures BTW.
From:
0
36
197
I got fluorescent plant extract. Glowing plant juice!
3
6
198
New tutorial on pathway visualization is live! The workflow models pathways as networks + visualize them using ggraph, a ggplot extension of network analysis package igraph. Tutorial, example scripts, input & output are available on GitHub:
3
38
194
@ImperfectFunGuy
Update: the
#PurplePastaSauce
microbe has been delivered to the microbiologist Dr.
@ImperfectFunGuy
! "I'm excited" is an understatement.
2
2
213
I got my “free” author copy. Technically it’s my boss’s copy, but she let me have it.
4
1
189
Wow! What an honor to be the author of one of the most starred R repos. Humbled to be along side with the repos for "R for Data Science" () and "Fundamentals of Data Visualization" ().
Most starred R repos:
- (
@hadleywickham
): 4219 stars
- (
@ChenxinLi2
): 3967 stars
- (
@ClausWilke
): 2999 stars
1
15
57
5
15
187
My "Simple Tidy Gene Co-expression" analysis workflow has been updated to include a 2nd use case - a heat stress time course. The workflow seems to function fine both to study development and stress. Read more:
5
34
187
Notice of Award! I will be working as postdoc/co-PI (an upgrade, I guess) on our newly funded NSF-EAGER to pilot single cell multiome and TF-DNA interaction assays in medicinal plants with
@TheBuellLab
. 🌳🌿🧬🧪💻📊
19
6
184
Benchmarked my gene co-expression analysis method against WGCNA. Both can detect early vs late development co-expression modules, but WGCNA returned more modules + higher spatial resolution (left). However, my method return tighter modules (right).
3
27
177
@chorye
I have an illustrated repository called “friends don’t let friends make bad graphs”.
Comes with counter-examples, examples, and explanations.
1
35
179
My "Friends don't like friends make bad graphs" repository reached 800 stars! Thanks everyone.
3
32
173
I realized I haven't announced it here yet: I'll be moving to Athens GA to join Dr Robin Buell's new lab at University of Georgia as a postdoc. I'll continue to work on plant (epi)genomics there.
12
2
172
I understand why people hate the X branding/its owner & want to jump ship to other apps. But I really hope we don’t disband science twitter. At conferences I use “I am Li, I follow you on Twitter” as ice breaker. It will stop working if we disperse ourselves to 5 different apps.
9
8
173
Now that I have reviewed for both Plant Cell and Plant Physiology, one thing I don’t like is that they ask the reviewer to estimate the impact or number of citations per year for the manuscript. Society journals shouldn’t be doing this.
19
12
170