Changes between Version 1 and Version 2 of 2024-8-8


Ignore:
Timestamp:
Aug 13, 2024, 10:40:04 AM (14 months ago)
Author:
Tom Goddard
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • 2024-8-8

    v1 v2  
    1111= Discussion Notes =
    1212
     13* SSD drives on plato
     14  - Each of 4 plato nodes now has a 4 TB SSD drive mounted at /scratch.
     15  - Scooter is using this to build BLAST databases with makeblastdb which fails to run on beegfs file system.
     16
     17* Martin Steinegger talk
     18  - Martin created BFD (big fantastic database) metagenomics database that DeepMind then used for AlphaFold 2.
     19  - Trained network to run Foldseek directly on a sequence with the network directly translating sequence to 3Di spatial alphabet.  This takes seconds instead of minutes to make a ColabFold prediction.
     20  - Martin has clustered AlphaFold database of 214 million structures down to 2-3 million clusters using foldseek-clust.
     21  - About 35% of clusters have no annotations.
     22  - Looked at how many clusters are specialized to bacteria or eukaryotes or archaea, cluster.foldseek.com.
     23  - Martin made pitch strongly advocating open source software.  Said they made foldseek open source in 2021 3 years before publication.
     24
    1325= Action Items =