lichtenstein.bsky.social
Biophysical chemist and Protein Designer. I like making new molecules do new things.
Group leader at the Centre for Enzyme Innovation at the University of Portsmouth.
22 posts
163 followers
41 following
Getting Started
Conversation Starter
comment in response to
post
Internationally oriented grants are not bad as your team offers a unique value added element; but have you been targeted your applications around tool development or a scientific question?
comment in response to
post
I have had representatives from companies ask me how well some of the more common algorithms work as a user; and while I can generally be positive, it’s always with a caveat — you need to understand how to read/edit code, there are unpublished best practices, their capabilities are context dependent
comment in response to
post
They do in general — the deviations in the cities are often related to forced foot traffic flows…
comment in response to
post
Always loved this! How long did it take you to find the music to accompany it?
comment in response to
post
So all proteins are in a superposition of folded and unfolded states; sounds about right actually.
comment in response to
post
This is going to be the death of the Europe-to-US conference attendance. I’m not sure Universities would insure trips where targeted deportation is a risk.
comment in response to
post
Scientific societies are in general fairly conservative entities, chemists are in general the most conservative scientists, so it doesn’t surprise me that academic societies of chemists can be regressive.
comment in response to
post
Hello, I’m a biophysical chemist and protein designer at the University of Portsmouth — could I be added to the science feed?
www.port.ac.uk/about-us/str...
comment in response to
post
The manuscript has a bit for everyone: manual protein design to preserve enzymatic thermodynamic properties with confirmatory structure, careful activity measurements, pilot scale bioprocessing of post-consumer plastic, synergy and more!
comment in response to
post
There are reasonable measures of information content for proteins of various sizes (and this is ultimately what we’re talking about, right?) — but as a rule of thumb if profile-profile HMM find natural proteins from your sequence, it’s probably not really de novo, regardless of percent similarities.
comment in response to
post
I only use it in the context of proteins designed without sequence relationship to naturally occurring proteins. But the accepted definition seems to be between that and the introduction of non-natural function into proteins. Certainly hand crafted proteins count?
comment in response to
post
There is almost certainly a bond rotation
comment in response to
post
Other classic examples are sugar pucker in DNA, and as others have pointed out cis-trans isomerisation of proline. Students never seem overly impressed with these details…
comment in response to
post
Would be great to be added!
comment in response to
post
A form of reflective selection bias — the algorithms train on proteins that expressed sufficiently to yield experimental structures, they end up with higher confidence in sequences/structures they have seen which affords proteins that express well.
comment in response to
post
Or perhaps simply the accuracy of the calculation? Force fields are ultimately approximations.
comment in response to
post
Hiya!