Systems Biologist trying to untangle Host-Drug-Microbiome interactions | PI at
@UMassMed
School | Trying to keep the BS to a minimum (not always succeeding)
Why are some non-antibiotic drugs toxic to bacteria 💊☠️🦠? Do they work like standard
#antibiotics
(ABX)? Our latest work by
@MarianaNoto
published
@ScienceMagazine
addresses this fundamental question (1/n)
I keep on preaching to my lab about the usefulness
@ChatGPTapp
for coding, but real-life demos might be more effective. A thread on a neural network we quickly developed to classify antibiotics sensitivity 🧵 (1/10)
That was a tough one. There were ~2,500 colonies, most humans and machines considerably underestimated. Follow the thread for details and anecdotes ... 🧵
Finally out on
@eLife
! We show that bacterial evolved resistance against chemotherapy can reduce its efficacy in a treated host. Potential important implications for the
#Cancer
-
#Microbiome
field. By
@BrittanyRosener
and
@serkansayin_
& the Walhout lab
Our paper on cell-cell variation in drug resistance in isogenic melanoma cells is out. Congrats to
@PayamKhoshkenar
and
@emily_lowry23
. A quick recap of our findings 🧵🧵🧵
With all the hype around single-cell/molecule/whatever methods, I’m still in awe by the simplicity and brilliance of
#OldSchool
assays like x-gal plates (though this specific one might be extremely important so I’m biased)
Final thoughts: 🤯 Great experience, blow my mind, ⏱️ saved us tons of time (took ~3hr), 👨💻 Will gladly share code/images if useful, 🔵 I started on the Blue place too, follow me there (10/10)
First publication of work done in our lab (
#microscopy
and auto. image analysis by
@PayamKhoshkenar
). Research was lead by the
@london_lab
.
Covalent Docking Identifies a Potent and Selective MKK7 Inhibitor.
@AnnaTesten
Why not sequence (WGS is 80$ with
@MiGScenter
)? I still have to pinch myself every time I realize that you can sequence 5 bacterial genomes for the price of a single antibody
@AMartinezArias
I rarely get mixed in these (don’t know any of you), but setting the science aside, I think your writing was overly aggressive (and continues to be so in your reply). This reflects a toxic vide many ppl dislike and it does a bad service to the wider scientific community.
We think confetti strains are a unique tools for educators and we actually planning a new
#STEM
#ScienceOutreach
program to follow our previous “Emergence of a superbug” project.
Necessity is the mother of
#creativity
. My student Alex taped a magnifying glass to our tissue hood screen to see the
#cancer
spheroids he’s working on. Works great!
Together with more than 150 peers, I just (zoom) attended one of the *best* PhD defences I ever saw. Remarkable delivery under extrodinary circumstances by
@CruzgordilloP
from the
@TheLeeLab
.
This time around, we’ll publish the results only Monday. And, to be completely honest, we got highly different estimates from our
@MATLAB
code,
@imagej
assisted counting and a manual physical counter.
A
@UMassMedical
Now piece on our
#STEM
outreach program with 250 high-school students. Our online
#robotic
platform allows remote students across 7,500 miles to evolve bacteria in our lab and identify mutations driving
#antibiotics
resistance.
Our joint opinion piece in
@trendscancer
following a fantastic
@NCIsysbio
and Xavier’s lab (
@sloan_kettering
) organized meeting. Cancer-Microbiome interactions are a new frontier for Systems Biology.
Very impressive study monitoring a defined
#microbione
in mice over 6-years!! Reveals adapation mechanisms and strain co-existance (sort of a
@RELenski
type
#LTEE
with a 12-species community and in-vivo).
Our lab, joining all labs in my department
@UMassMedical
, has contributed all of our
#PPE
(personal protection equitment) to
@umassmemorial
. Glad we can help. Thanks to "our boots on the ground” still on campus
@serkansayin_
, Cedric Diot, Nicki Fox and
@YuLiu_Sunny
Congrats
@asanchez_lab
for really elegant work! Assembling *all* possible permutations of N microbes (with a multi-pipette) and evaluate their performance (they explore N=8, P. aeruginosa).
That’s it! We have a working NN classifier!! Next, we wrote a wrapper script that takes in new plate images and uses the trained neural network for classification of each well. It worked great!!! (9/10)
@Prokaryota
I often tell my students that debugging failed PCRs is one of the best wax-on/wax-off exercises for biologists. Unfortunately, this reference typically fails to cross the
#GenerationGap
(as most of my other pop-culture refs)
Brittany (
#NotOnTwitterYet
) and
@serkansayin_
from my lab are about to sequence genomes of evolved bacteria with
#MiniSeq
(thanks to
@athmapai
). This is our equivalent to "boldly go where no one has gone before”. Unbelievable that we’ll get to explore these tomorrow!!!
Lastly, this paper is also the outcome of a very professional editorial process that truly added value to the process, pushed against unreasonable requests of the “reviewer
#2
” and helped us focus. Past tweets by others tell me many share this experience at
@CellSystemsCP
Many people are behind this success, but this is primarily the triumph 🏆 of
@MarianaNoto
who confronted a challenge that seemed truly intractable (study the MOA of almost 200 drugs) (18/19)
We used a microplate reader to grow and measure absorbance over 16hr. Strains were diluted 1:200 from OD 1 into LB (rich media) or M9 (synthetic media) supplemented with kanamycin.
In Jan 2021, I'll be moving to
@Umassmed
to join the Program in Systems Biology (PSB) as associate prof. Superb place w/ great labs incl
@job_dekker
@mitchell_sysbio
@TheLeeLab
Walhout, Brewster, Shank +more. I thank PSB for the opportunity & my students for making this possible.
A very thoughtful perspective on Feedback in biological systems from
@HanaScientist
on
@CellSystemsCP
. Begins with what we know, continues with key unanswered questions and has a beautiful poetic ending!
The Golub lab (
@broadinstitute
) stares the elephant in the room straight in the eyes: Inescapable rapid evolution of cancer cell-lines across labs leads to diverse phenotypes and drug sensitivities.
We first made a simple plastic adapter that allowed us to quickly image microdilution plates with a standard document scanner (keeping plates in the same position) (4/10)
We manually divided the single-well images into subfolders by their phenotypes. Nothing fancy, drag and drop according to a decision of a human classifier (6/10)
Brittany,
@serkansayin_
and our
@SingerInst
robot rocking a genetic screen after a long time. Always exciting to see high-throughput biology being made in a push of a button.
@blekhman
I see it as a kind of a “statement", like hipsters using a flip phones to show they aren’t swayed by technology or using a Mac to signal you’re an individual thinker
We wrote a
@matlab
script to segment plate photos, crop them, and save images of individual wells (by position). Scanning 40 plates generated a our training set (~4000 images) (5/10)
Most labs score growth by eye, but that seems overly *tedious*. Since we have hundreds of plates to test, and a crave for tech, we decided to try out an
#AI
solution (3/10)
Just covered this paper in our journal club. Beautiful work by Erdal Topeka’s lab (
@UTSWNews
) on bacterial adaptation to antibiotics in mice GI tract (without evolving resistance)
Our
#STEM
project ends with a bang. Whole-Genome-Sequencing (
#WGS
) of an evolved bacterial strain homes resistance to an inactivating mutation in ompR. Special thanks to the 200 world-wide
#human
participants (and one
@Eppendorf_USA
#robot
)!
We then used
@MATLAB
to fit a growth curve function and extract key parameters: doubling time, lag period, and yield (3 biological replicates per strain/condition)
Drug microdilution panels are a common method for testing bacterial
#antibiotic
sensitivity. Basically, you aliquot bacteria into multi-well plates with a different drugs and monitor growth after a day (2/10)
Welcoming Peter for a rotation in our lab (Brittany
#NotOnTwitterYet
is showing him around). He’ll be developing a method for bacterial genetic screens using
@illumina
#NGS
When added results from screen of non-ABX to the mix ▲◼️, almost all of them turned out to be disconnected from the ABX hubs. This was a key observation 🔑: non-ABX likely kill bacteria differently from standard ABX (10/19)
Our lab is celebrating. Esther joining a summer project, Brittany joining the lab, myself for grant submission,
@serkansayin_
for his qualifying exam and
@PayamKhoshkenar
for his contribution to a manuscript with collaborators.
If you’re interested in drug resistance (across fields), consider attending the Gordon conference
@kaushans
and I are chairing this June (we have spots for contributed talks)
Over-estimation of stress-induced mutagenesis puts evolvability in question (in bacteria). Great insights *simple* and beautiful experiments. Way to go
@PLOSBiology
. CNS needs to publish more of these!
Data points for yield should be taken with a grain of salt since the colors of the chromoproteins might impact absorbance (especially at high confluence).
Just added another
@ProgMolMed
from
@UMassMedical
to our
@DataCamp
training on
#DataScience
. Almost 80 trainees, and some of them are racing forward. I strongly recommend Bio departments world-wide to consider this for remote training for wet and dry bio students.
Next, we prompted chatGPT to generate
@MATLAB
script for a neural-network that classifies hand-written digits (an input format is very similar to our single-well photos) (7/10)
Tour-de-force detective work on cellular memory in yeasts by
@KevinVerstrepen
and
@swainqbiolab
. *Beautifully* written and leaving room for a sequel work. Glad to see that
@eLife
gets that not *every* paper in cell bio. needs to end with a mechanism.
We modified the code slightly (e.g., we have only 8 alternative classes while digits have 10) and trained the neural network on our training set (8/10)
Finally the S. cerevisiae strikes back. Still a few K generations behind
@RELenski
's LTEE but “a journey of a thousand miles begins with a single step"
The longest evolution experiment in yeast is still going strong! It confirms findings from the famous long-term evolution experiment in E. coli and identifies interesting differences. Retweet if your network should learn about this window into evolution
3 years ago my lab had its first taste of
#NGS
. Since then our hunger for sequencing doubled every 6 months. As a postdoc I swore by quantitative microscopy … as a PI I feel that my journey towards the dark side is now complete ⚫️
Our brief
@RELenski
#outreach
#evolution
project ended. Results from 96 adapted strains (evolved by 150 high-schoolers across the Atlantic accessing our online robotic platform).
(resistance inferred from fifty plates like this one)
@DrHazmatic
🚶Walking station desk, sounds absurd, but gets you moving daily and 🖍 Crayons, they work on white boards and are erased w/o a mark with a wet cloth (even after 1yr)