I am once again asking big name professors to give a talk on 1-2 projects and not high level overviews of 20
| Google Scholar | https://scholar.google.com/citations?user=yfARKJgAAAAJ&hl=en |
James Kath
- 86 Followers
- 175 Following
- 51 Posts
Chicago-area chemical biologist. Previously Northwestern (protein glycosylation) and Harvard (DNA replication/repair). Formally @rikeijames. Personal account.
If I wanted to just look at graphical abstracts of your CNS papers, I could do that on my own time
Any research article/review recommendations if I want to catch up on what’s current in protein design (AKA the David Baker Extended Universe)?
Carl ZimmerTom Lehrer is still alive [!] And he put his work in the public domain. https://tomlehrersongs.com/
Keith HornbergerIn case you haven’t been keeping up with the American Chemical Society shit fiesta this week, they fired two editors at Chemical & Engineering News, including the EIC. This milquetoast editorial full of meaningless corporate platitudes basically says they weren’t getting with the program of turning CE&N from a fine chemistry journalism magazine into a shill for the ACS. Not renewing my 25+ year membership, and I don’t do such things lightly.
https://cen.acs.org/acs-news/Editorial-Change/100/i44
https://cen.acs.org/acs-news/Editorial-Change/100/i44
Any research article/review recommendations if I want to catch up on what’s current in protein design (AKA the David Baker Extended Universe)?
Michael BaymMaybe the real light of evolution was the ad-hoc explanations we made along the way
Helen BranswellMy colleague, Nick Florko, spent three years reporting this tour de force project on how state prisons are allowing inmates to die of a curable disease, #hepatitisC. It is so worth your time, though be forewarned: It is heartbreaking & infuriating. https://www.statnews.com/death-sentence/
2) What other possible applications could there be, and when would they be most useful in the development of antibody-based therapeutics.
One that jumps out to me is a similar display/sequencing-based A vs. B (or vs. C/D/E) screening of a target vs. anti-target(s). It seems another recent paper they cite, moves more in that direction.
https://www.sciencedirect.com/science/article/pii/S2667237522001278
In addition to uploading code, the authors also posted a web-based interface where you can input your own nanobody sequences.
http://18.224.60.30:3000/
What I'm really interested in going forward is:
1) How much data (sequencing-based or otherwise) would you need to get this degree of performance from a logistic model? They used ~60,000 sequences for "high" and "low" PSR binding (before some filtering), and later find if you increase it to 1M for each, the neural networks do a little better.