Back

Gene symbol recognition with GeneOCR

Geser, P.; Zhang, J. D.

2022-07-03 bioinformatics
10.1101/2022.07.01.498459 bioRxiv
Show abstract

MotivationFigures in biomedical publications often contain lists of gene symbols. Human intelligence and manual curation is required to recognize, extract, and validate the symbols in order to reuse them. The task is laborious, error prone, and amenable to automation. ResultsWe introduce GeneOCR, a web-based software to recognize, extract, and check gene symbols from figures. It features both an interactive user interface and programmatic access. Simulation-based benchmark reveals that GeneOCR recognizes or suggests correct gene symbols in more than 80% cases. The performance qualifies GeneOCR as an enhancement to manual curation. AvailabilityThe source code of GeneOCR is available with GNU v3 license at https://github.com/bedapub/GeneOCR-UI (frontend) and https://github.com/bedapub/GeneOCR-API (backend). A Docker/Singularity image is available via BioContainers. Contactjitao_david.zhang@roche.com

Matching journals

The top 2 journals account for 50% of the predicted probability mass.

1
Bioinformatics
◐ 1204 papers in training set
Top 0.4%
34.4%
Similar papers in this journal
2
BMC Bioinformatics
○ 457 papers in training set
Top 0.2%
18.6%
Similar papers in this journal
50% of probability mass above
3
Bioinformatics Advances
◐ 203 papers in training set
Top 0.3%
8.9%
Similar papers in this journal
4
GigaScience
◐ 212 papers in training set
Top 0.5%
5.5%
Similar papers in this journal
5
PLOS ONE
● 5266 papers in training set
Top 29%
4.9%
Similar papers in this journal
6
Computational and Structural Biotechnology Journal
● 242 papers in training set
Top 3%
2.0%
7
Nucleic Acids Research
◐ 1281 papers in training set
Top 9%
1.7%
8
PeerJ
◐ 308 papers in training set
Top 6%
1.4%
9
SoftwareX
○ 15 papers in training set
Top 0.2%
1.1%
10
Journal of Open Source Software
● 25 papers in training set
Top 0.3%
1.1%
11
NAR Genomics and Bioinformatics
◐ 242 papers in training set
Top 3%
1.1%
12
PLOS Computational Biology
● 1863 papers in training set
Top 17%
1.1%
13
Computer Methods and Programs in Biomedicine
○ 28 papers in training set
Top 0.8%
1.0%
14
BioData Mining
○ 22 papers in training set
Top 0.5%
1.0%
15
Briefings in Bioinformatics
◐ 354 papers in training set
Top 7%
0.8%
16
Gigabyte
● 62 papers in training set
Top 1%
0.8%
17
Journal of Bioinformatics and Systems Biology
○ 15 papers in training set
Top 0.3%
0.8%
18
Scientific Reports
○ 3612 papers in training set
Top 72%
0.8%
19
Database
◐ 61 papers in training set
Top 0.9%
0.8%

"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.