Baffling 🤯 Here’s to hoping that time and resources are increasingly given to support widely-used software that underlies modern medical research. Having dealt with this issue, and others very similar, I can only hope to express to you all how frustrating this can be
(1/11) Genomics databases, like the SRA, host tons of data. But data access is challenging, driven largely by tooling. Below is a brief history on genomics data access. 🧵
A new showing that even when AlphaFold 2 structures differ from the crystal structure, the predictions are often telling us something true about physical reality. You can dock compounds in them with high hit rates and the cryoEM shows the AF2 structure, not the crystal!
@theHumanBorch Will definitely have to pick your brain once I’m able to read through it. I know of another project with a good amount of VDJ seq data this could help with
Can’t wait to see this paper! One of the biggest challenges with this data is being able to predict antigen specificity. 5/5 sounds pretty great to me 🎯
Shout out to @LabWaggoner for constantly sharing cutting edge work, and big thank you to whoever or whatever algorithm runs this account!
If you’re looking for a high quality feed of science, give them a follow
@davisidarta Can’t wait to try this out. Any comment on how or if this would work with samples that need integrating? I would think that without integration/dim reduction, many of the subpopulation specific markers may be due to artifact or technology used.
@Jeff_Mold @ShainLab Sorry, I don’t do spatial at the moment. I would agree that spatial is a bit harder to judge and one would expect a bit of variation due to differences in cell types across locations. Maybe do multiple known samples to set cutoffs and go from there
@UICancerBiology @UIowaCancer@Iowa_MSTP Thank you! Couldn’t have done it without the great students in the program.
Can’t wait to dive deeper into CTCL with the help of my committee