Back

PDBCleanV2: A Python Library for Generating Consistent Structure Datasets

Pardo-Avila, F.; Weiner, L.; Cabral, P.; Corsepius, N. C.; Poitevin, F.; Levitt, M.

2025-02-19 bioinformatics
10.1101/2025.02.14.638326 bioRxiv
Show abstract

The number of structures in the Protein Data Bank has grown rapidly in recent years. Nonetheless, comparing structures of a specific system often proves challenging due to variant nomenclature, origins from different species, presence of various ligands, and missing atoms or chains. To address these issues, we have developed PDBCleanV2, a Python package that enables users to create consistent datasets of structures, simplifying comparison among structures. This library generates individual files for each molecule in a structure file, corrects labeling errors, and standardizes chain names and numbering. Our aim is to provide researchers with consistent datasets that streamline their analysis. The source code, installation instructions, and tutorials can be found in PDBCleanV2s GitHub repository and Zenodo accessible at https://github.com/fatipardo/PDBClean-0.0.2/ and https://doi.org/10.5281/zenodo.14014241, respectively.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

1
Protein Science
○ 246 papers in training set
Top 0.1%
22.0%
Similar papers in this journal
2
Bioinformatics
◐ 1204 papers in training set
Top 3%
9.5%
Similar papers in this journal
3
Journal of Chemical Information and Modeling
● 238 papers in training set
Top 0.9%
5.4%
Similar papers in this journal
4
Nature Communications
○ 5641 papers in training set
Top 29%
4.8%
Similar papers in this journal
5
Journal of Molecular Biology
○ 232 papers in training set
Top 0.6%
4.3%
Similar papers in this journal
6
Nucleic Acids Research
◐ 1281 papers in training set
Top 5%
4.0%
7
Nature Methods
○ 385 papers in training set
Top 3%
4.0%
50% of probability mass above
8
Bioinformatics Advances
◐ 203 papers in training set
Top 2%
3.5%
9
Nature Protocols
○ 33 papers in training set
Top 0.1%
3.5%
10
PLOS Computational Biology
● 1863 papers in training set
Top 10%
3.2%
11
Structure
○ 193 papers in training set
Top 0.9%
2.7%
12
Acta Crystallographica Section D Structural Biology
● 59 papers in training set
Top 0.2%
2.4%
13
Briefings in Bioinformatics
◐ 354 papers in training set
Top 4%
2.4%
14
Proteins: Structure, Function, and Bioinformatics
○ 88 papers in training set
Top 0.7%
1.9%
15
Proceedings of the National Academy of Sciences
● 2444 papers in training set
Top 28%
1.7%
16
Communications Biology
○ 993 papers in training set
Top 14%
1.7%
17
Molecular Biology and Evolution
◐ 542 papers in training set
Top 3%
1.7%
18
Journal of Cheminformatics
○ 29 papers in training set
Top 0.4%
1.7%
19
eLife
● 5828 papers in training set
Top 54%
1.5%
20
Scientific Data
○ 209 papers in training set
Top 2%
1.5%
21
BMC Bioinformatics
○ 457 papers in training set
Top 4%
1.3%
22
Computational and Structural Biotechnology Journal
● 242 papers in training set
Top 5%
1.1%
23
Scientific Reports
○ 3612 papers in training set
Top 68%
1.0%
24
Database
◐ 61 papers in training set
Top 0.8%
1.0%
25
Journal of Structural Biology
○ 64 papers in training set
Top 0.6%
1.0%
26
F1000Research
○ 88 papers in training set
Top 5%
0.6%
27
GigaScience
◐ 212 papers in training set
Top 5%
0.6%
28
Journal of Open Source Software
● 25 papers in training set
Top 0.5%
0.6%
29
IUCrJ
● 32 papers in training set
Top 0.3%
0.6%

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.