Upload README.md
Browse files
README.md
CHANGED
@@ -1,40 +1,109 @@
|
|
1 |
-
---
|
2 |
-
license: cc-by-nc-4.0
|
3 |
-
dataset_info:
|
4 |
-
features:
|
5 |
-
|
6 |
-
|
7 |
-
|
8 |
-
|
9 |
-
|
10 |
-
|
11 |
-
|
12 |
-
|
13 |
-
|
14 |
-
|
15 |
-
|
16 |
-
|
17 |
-
|
18 |
-
|
19 |
-
splits:
|
20 |
-
|
21 |
-
|
22 |
-
|
23 |
-
|
24 |
-
|
25 |
-
|
26 |
-
|
27 |
-
|
28 |
-
|
29 |
-
download_size: 134650306
|
30 |
-
dataset_size: 175825691
|
31 |
-
configs:
|
32 |
-
- config_name: default
|
33 |
-
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: cc-by-nc-4.0
|
3 |
+
dataset_info:
|
4 |
+
features:
|
5 |
+
- name: accession
|
6 |
+
dtype: string
|
7 |
+
- name: name
|
8 |
+
dtype: string
|
9 |
+
- name: Full Name
|
10 |
+
dtype: string
|
11 |
+
- name: taxon
|
12 |
+
dtype: string
|
13 |
+
- name: sequence
|
14 |
+
dtype: string
|
15 |
+
- name: function
|
16 |
+
dtype: string
|
17 |
+
- name: AlphaFoldDB
|
18 |
+
dtype: string
|
19 |
+
splits:
|
20 |
+
- name: train
|
21 |
+
num_bytes: 168911398
|
22 |
+
num_examples: 248315
|
23 |
+
- name: test
|
24 |
+
num_bytes: 3470418
|
25 |
+
num_examples: 4203
|
26 |
+
- name: validation
|
27 |
+
num_bytes: 3443875
|
28 |
+
num_examples: 4172
|
29 |
+
download_size: 134650306
|
30 |
+
dataset_size: 175825691
|
31 |
+
configs:
|
32 |
+
- config_name: default
|
33 |
+
data_files:
|
34 |
+
- split: train
|
35 |
+
path: data/train-*
|
36 |
+
- split: test
|
37 |
+
path: data/test-*
|
38 |
+
- split: validation
|
39 |
+
path: data/validation-*
|
40 |
+
---
|
41 |
+
|
42 |
+
|
43 |
+
|
44 |
+
# Dataset Card for Prot2Text-Data
|
45 |
+
|
46 |
+
|
47 |
+
## Dataset Description
|
48 |
+
|
49 |
+
This dataset is designed for training the Prot2Text framework. It contains information for 256,690 proteins and is comprised of three modalities:
|
50 |
+
|
51 |
+
Protein sequence (amino acid sequence)
|
52 |
+
AlphaFold accession ID
|
53 |
+
Textual description of the protein
|
54 |
+
The dataset is built from the SwissProt database, a component of UniProtKB Release 2022_04.
|
55 |
+
|
56 |
+
## Dataset Summary
|
57 |
+
|
58 |
+
Format: TSV (Tab-separated values) - expected (not explicitly mentioned)
|
59 |
+
Splits: Train (248,315), Validation (4,172), Test (4,203)
|
60 |
+
Size: 256,690 proteins
|
61 |
+
Features:
|
62 |
+
name (str): Full name of the protein
|
63 |
+
sequence (str): Amino acid sequence of the protein
|
64 |
+
AlphaFoldDB (str): Accession ID of the protein in AlphaFold database
|
65 |
+
taxon (str): Species information (potentially)
|
66 |
+
text (str): Textual description of the protein (excluding "(By Similarity)" and "PubMed")
|
67 |
+
## Supported Tasks and Leaderboards
|
68 |
+
|
69 |
+
This dataset is intended for training and evaluating models that learn relationships between protein sequences, structures, and textual descriptions. Tasks may include:
|
70 |
+
|
71 |
+
Protein function prediction
|
72 |
+
Protein structure prediction
|
73 |
+
Text-based protein information retrieval
|
74 |
+
## Dataset Structure
|
75 |
+
|
76 |
+
#### Data Fields:
|
77 |
+
|
78 |
+
* **name:** (string) codename of the protein.
|
79 |
+
* **Full Name:** (string) Full name of the protein.
|
80 |
+
* **sequence:** (string) Amino acid sequence of the protein.
|
81 |
+
* **AlphaFoldDB:** (string) Accession ID of the protein in AlphaFold database.
|
82 |
+
* **taxon:** (string) Species information for the protein.
|
83 |
+
* **function:** (string) Textual description of the protein.
|
84 |
+
|
85 |
+
|
86 |
+
#### Data Splits:
|
87 |
+
|
88 |
+
The dataset is split into training, validation, and test sets with a maximum sequence similarity threshold of 40% within each set using the CD-HIT clustering algorithm.
|
89 |
+
|
90 |
+
* **Train:** 248,315 proteins
|
91 |
+
* **Validation:** 4,172 proteins
|
92 |
+
* **Test:** 4,203 proteins
|
93 |
+
|
94 |
+
## Considerations for Using the Data
|
95 |
+
|
96 |
+
The dataset is built from a single source (SwissProt). Consider incorporating data from other sources to increase diversity.
|
97 |
+
The textual descriptions may contain biases present in the original database. Be mindful of these biases when using the data for downstream tasks.
|
98 |
+
|
99 |
+
|
100 |
+
### License
|
101 |
+
We are releasing this dataset under the terms of [CC-BY-NC-4.0](https://opendatacommons.org/licenses/by/1-0/).
|
102 |
+
|
103 |
+
|
104 |
+
### Citation
|
105 |
+
Please cite this dataset and the original sources if you use this dataset in your work
|
106 |
+
```none
|
107 |
+
|
108 |
+
```
|
109 |
+
|