Published March 9, 2025 | Version v2
Dataset Open

Investigating the performance of foundation models on human 3'UTR sequences - data

  • 1. ROR icon Helmholtz Zentrum München
  • 2. ROR icon Technical University of Munich

Description

Data accompanying the publication Sergey Vilov and Matthias Heinig "Investigating the performance of foundation models on human 3'UTR sequences"

This data can be analyzed with our associated Github code: https://github.com/sergeyvilov/investigating-foundation-models-3utr

The model weigths for 3'UTR-specific models can be found in models/

The raw scores of all models on all tasks can be found in all_scores/

Files

all_scores.zip

Files (21.1 GB)

Name Size Download all
md5:aeb7ffae5ee4f293e0789f776e03acb1
181.7 MB Preview Download
md5:f832ede0bdd490905d4a23779197fe25
2.5 GB Preview Download
md5:c08e1659de671c0bce9103df8f330fd3
331.2 MB Preview Download
md5:b8e12c3c81da0d9b9e1f0663ab29fea2
7.8 GB Preview Download
md5:19f9507e97fd99f66116584e4eb7cf8f
4.7 GB Preview Download
md5:b9e2bde1df65e9b057d6021892af4cde
4.3 MB Preview Download
md5:d0b32bfeb31b9901edb24a6aaad3e14f
2.1 GB Preview Download
md5:178fe02dbbe43085499d86092a8c932f
122.5 MB Preview Download
md5:f626750489eae2ae038067479a1b4f5b
3.4 GB Preview Download

Additional details

Related works

Is part of
Preprint: 10.1101/2024.02.09.579631 (DOI)