Spaces:

PascalNotin
/

Tranception_design

Build error

App Files Files Community

PascalNotin commited on Aug 25, 2022

Commit

8eb33a8

•

1 Parent(s): b07be69

Fixed typo

Browse files

Files changed (1) hide show

app.py +1 -1

app.py CHANGED Viewed

@@ -111,7 +111,7 @@ title = "Interactive in silico directed evolution with Tranception"
 description = "Perform in silico directed evolution with Tranception to iteratively improve the fitness of a starting protein sequence one mutation at a time. At each step, the Tranception model computes the log likelihood ratios of all possible single amino acid substitution Vs the starting sequence, and outputs a fitness heatmap and recommandations to guide the selection of the mutation to apply. Note: The current version does not currently leverage homologs retrieval at inference time to boost fitness prediction performance."
 article = "<p style='text-align: left'>**Tranception: Protein Fitness Prediction with Autoregressive Transformers and Inference-time Retrieval**</p>"
 article += "<br><p style='text-align: left'> Pascal Notin, Mafalda Dias, Jonathan Frazer, Javier Marchena-Hurtado, Aidan N. Gomez, Debora S. Marks<sup>*</sup>, Yarin Gal<sup>*</sup>"
-article += "<br><p style='text-align: left'> <a href='https://proceedings.mlr.press/v162/notin22a.html' target='_blank'>Paper</a> *** <a href='https://github.com/OATML-Markslab/Tranception' target='_blank'>Code</a> </p>
 examples=[
 ['ADRB2_HUMAN --> MGQPGNGSAFLLAPNGSHAPDHDVTQERDEVWVVGMGIVMSLIVLAIVFGNVLVITAIAKFERLQTVTNYFITSLACADLVMGLAVVPFGAAHILMKMWTFGNFWCEFWTSIDVLCVTASIETLCVIAVDRYFAITSPFKYQSLLTKNKARVIILMVWIVSGLTSFLPIQMHWYRATHQEAINCYANETCCDFFTNQAYAIASSIVSFYVPLVIMVFVYSRVFQEAKRQLQKIDKSEGRFHVQNLSQVEQDGRTGHGLRRSSKFCLKEHKALKTLGIIMGTFTLCWLPFFIVNIVHVIQDNLIRKEVYILLNWIGYVNSGFNPLIYCRSPDFRIAFQELLCLRRSSLKAYGNGYSSNGNTGEQSGYHVEQEKENKLLCEDLPGTEDFVGHQGTVPSDNIDSQGRNCSTNDSLL', 1, 10, "Small", True],
 ['IF1_ECOLI --> MAKEDNIEMQGTVLETLPNTMFRVELENGHVVTAHISGKMRKNYIRILTGDKVTVELTPYDLSKGRIVFRSR', 1, None, "Medium", False],

 description = "Perform in silico directed evolution with Tranception to iteratively improve the fitness of a starting protein sequence one mutation at a time. At each step, the Tranception model computes the log likelihood ratios of all possible single amino acid substitution Vs the starting sequence, and outputs a fitness heatmap and recommandations to guide the selection of the mutation to apply. Note: The current version does not currently leverage homologs retrieval at inference time to boost fitness prediction performance."
 article = "<p style='text-align: left'>**Tranception: Protein Fitness Prediction with Autoregressive Transformers and Inference-time Retrieval**</p>"
 article += "<br><p style='text-align: left'> Pascal Notin, Mafalda Dias, Jonathan Frazer, Javier Marchena-Hurtado, Aidan N. Gomez, Debora S. Marks<sup>*</sup>, Yarin Gal<sup>*</sup>"
+article += "<br><p style='text-align: left'> <a href='https://proceedings.mlr.press/v162/notin22a.html' target='_blank'>Paper</a> *** <a href='https://github.com/OATML-Markslab/Tranception' target='_blank'>Code</a> </p>"
 examples=[
 ['ADRB2_HUMAN --> MGQPGNGSAFLLAPNGSHAPDHDVTQERDEVWVVGMGIVMSLIVLAIVFGNVLVITAIAKFERLQTVTNYFITSLACADLVMGLAVVPFGAAHILMKMWTFGNFWCEFWTSIDVLCVTASIETLCVIAVDRYFAITSPFKYQSLLTKNKARVIILMVWIVSGLTSFLPIQMHWYRATHQEAINCYANETCCDFFTNQAYAIASSIVSFYVPLVIMVFVYSRVFQEAKRQLQKIDKSEGRFHVQNLSQVEQDGRTGHGLRRSSKFCLKEHKALKTLGIIMGTFTLCWLPFFIVNIVHVIQDNLIRKEVYILLNWIGYVNSGFNPLIYCRSPDFRIAFQELLCLRRSSLKAYGNGYSSNGNTGEQSGYHVEQEKENKLLCEDLPGTEDFVGHQGTVPSDNIDSQGRNCSTNDSLL', 1, 10, "Small", True],
 ['IF1_ECOLI --> MAKEDNIEMQGTVLETLPNTMFRVELENGHVVTAHISGKMRKNYIRILTGDKVTVELTPYDLSKGRIVFRSR', 1, None, "Medium", False],