Add semantic mining citation, update evo 1.5 config references (#105)

* Update README with citation

* Update acr_sample.json evo 1.5 name

* Update gene_completion.json evo 1.5 name

* Update operon_completion.json evo 1.5 name

* Update toxin_antitoxin_sample.json evo 1.5 name

Author: adititm

semantic_mining/README.md

@@ -74,4 +74,20 @@ Reference sequences for the operon and gene completion scripts are provided in t
 ## Related Documentation
 
 For more comprehensive documentation, please refer to:
-- The [Main Repository README](https://github.com/evo-design/evo/blob/main/README.md) for setting up generation with Evo
+- The [Main Repository README](https://github.com/evo-design/evo/blob/main/README.md) for setting up generation with Evo
+
+## Citation
+
+Please cite the following publication when referencing semantic mining or Evo 1.5.
+
+```
+@article {merchant2024semantic,
+   author = {Merchant, Aditi T and King, Samuel H and Nguyen, Eric and Hie, Brian L},
+   title = {Semantic mining of functional de novo genes from a genomic language model},
+   year = {2024},
+   doi = {10.1101/2024.12.17.628962},
+   publisher = {Cold Spring Harbor Laboratory},
+   URL = {https://www.biorxiv.org/content/early/2024/12/18/2024.12.17.628962},
+   journal = {bioRxiv}
+}
+```

semantic_mining/sample_configs/acr_sample.json

@@ -9,7 +9,7 @@
     "output_folds_file": "/semantic_mining/test/ouput_folds.csv",
     "output_filtered_folds": "/semantic_mining/test/filtered_folds.csv",
 
-    "model_name": "evo_1.5_8k_hf",
+    "model_name": "evo-1.5-8k-base",
     "n_tokens": 1000,
     "temperature": 0.7,
     "top_k": 4,

semantic_mining/sample_configs/gene_completion.json

@@ -13,7 +13,7 @@
     "segmasker_path": "/usr/local/bin/ncbi-blast-2.16.0+/bin/segmasker",
     "mafft_path": "/usr/local/bin/mafft/bin/mafft",
 
-    "model_name": "evo_1.5_8k_hf",
+    "model_name": "evo-1.5-8k-base",
     "n_tokens": 2500,
     "temperature": 0.7,
     "top_k": 4,

semantic_mining/sample_configs/operon_completion.json

@@ -13,7 +13,7 @@
     "segmasker_path": "/usr/local/bin/ncbi-blast-2.16.0+/bin/segmasker",
     "mafft_path": "/usr/local/bin/mafft/bin/mafft",
 
-    "model_name": "evo_1.5_8k_hf",
+    "model_name": "evo-1.5-8k-base",
     "n_tokens": 2500,
     "temperature": 0.7,
     "top_k": 4,

semantic_mining/sample_configs/toxin_antitoxin_sample.json

@@ -11,7 +11,7 @@
     "shared_protein_pairs_csv": "/semantic_mining/test/paired_candidates.csv", 
     "cofold_fasta": "/semantic_mining/test/paired_candidates.fasta", 
 
-    "model_name": "evo_1.5_8k_hf",
+    "model_name": "evo-1.5-8k-base",
     "n_tokens": 2000,
     "temperature": 0.7,
     "top_k": 4,