Update README.md
Browse files
README.md
CHANGED
@@ -32,3 +32,79 @@ configs:
|
|
32 |
- split: test
|
33 |
path: data/test-*
|
34 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
32 |
- split: test
|
33 |
path: data/test-*
|
34 |
---
|
35 |
+
|
36 |
+
# PPI: Protein-Protein Interfaces
|
37 |
+
|
38 |
+
|
39 |
+
## Overview
|
40 |
+
|
41 |
+
This task relates to predicting which pairs of amino acids, spanning two
|
42 |
+
different proteins, will interact upon binding (when they form a complex).
|
43 |
+
|
44 |
+
Amino acids are defined as interacting if any of their heavy atoms are within 6
|
45 |
+
Angstroms from one another.
|
46 |
+
|
47 |
+
|
48 |
+
## Datasets
|
49 |
+
- splits:
|
50 |
+
- DIPS-split: DIPS dataset, split by sequence identity (see add. inf.)
|
51 |
+
|
52 |
+
|
53 |
+
## Format
|
54 |
+
|
55 |
+
Each entry in the dataset contains the following keys:
|
56 |
+
|
57 |
+
|
58 |
+
['input_ids'] The set of atomic numbers for both the proteins in the complex concatenated together.
|
59 |
+
['coords'] The 3D coordinates for both the proteins in the complex concatenated together.
|
60 |
+
['label'] 1 for interacting, 0 for not interacting.
|
61 |
+
['token_type_ids'] A mask that corresponds to which inputs_ids/coords belong to the two individual proteins involved in the complex (0 for protein 1 , 1 for protein 2)
|
62 |
+
|
63 |
+
|
64 |
+
['atoms_pairs'] (pandas.DataFrame) Atom coordinates of the complexes.
|
65 |
+
['atoms_neighbors'] (pandas.DataFrame) Indicates which amino acids interacting in each complex.
|
66 |
+
['id'] (str) Contains the PDB filenames, model numbers, and chain IDs of the complexes
|
67 |
+
(format: <PDB_FILE>_<MODEL_NUM1>_<CHAIN_ID1>_<PDB_FILE>_<MODEL_NUM2>_<CHAIN_ID2>)
|
68 |
+
['types'] (dict) Type of each entry.
|
69 |
+
['file_path'] (str) Path to the LMDB.
|
70 |
+
|
71 |
+
|
72 |
+
## Additional Information
|
73 |
+
|
74 |
+
The ensemble consists of a given protein complex, with subunits corresponding
|
75 |
+
to the two individual proteins involved in the complex.
|
76 |
+
|
77 |
+
## Citation Information
|
78 |
+
|
79 |
+
```
|
80 |
+
@article{townshend2020atom3d,
|
81 |
+
title={Atom3d: Tasks on molecules in three dimensions},
|
82 |
+
author={Townshend, Raphael JL and V{\"o}gele, Martin and Suriana, Patricia and Derry, Alexander and Powers, Alexander and Laloudakis, Yianni and Balachandar, Sidhika and Jing, Bowen and Anderson, Brandon and Eismann, Stephan and others},
|
83 |
+
journal={arXiv preprint arXiv:2012.04035},
|
84 |
+
year={2020}
|
85 |
+
}
|
86 |
+
```
|
87 |
+
|
88 |
+
```
|
89 |
+
@article{paszke2019advances,
|
90 |
+
title={Advances in neural information processing systems 32},
|
91 |
+
author={Paszke, Adam and Gross, Sam and Massa, Francisco and Lerer, Adam and Bradbury, James and Chanan, Gregory and Killeen, Trevor and Lin, Zeming and Gimelshein, Natalia and Antiga, Luca and others},
|
92 |
+
journal={Curran Associates, Inc},
|
93 |
+
pages={8024--8035},
|
94 |
+
year={2019}
|
95 |
+
}
|
96 |
+
```
|
97 |
+
|
98 |
+
```
|
99 |
+
@article{vreven2015updates,
|
100 |
+
title={Updates to the integrated protein--protein interaction benchmarks: docking benchmark version 5 and affinity benchmark version 2},
|
101 |
+
author={Vreven, Thom and Moal, Iain H and Vangone, Anna and Pierce, Brian G and Kastritis, Panagiotis L and Torchala, Mieczyslaw and Chaleil, Raphael and Jim{\'e}nez-Garc{\'\i}a, Brian and Bates, Paul A and Fernandez-Recio, Juan and others},
|
102 |
+
journal={Journal of molecular biology},
|
103 |
+
volume={427},
|
104 |
+
number={19},
|
105 |
+
pages={3031--3041},
|
106 |
+
year={2015},
|
107 |
+
publisher={Elsevier}
|
108 |
+
}
|
109 |
+
```
|
110 |
+
|