Extract and normalize variants from academic papers in xml, pdf, doc, docx, xlsx, csv formats.
- Linux OS
- Docker 18.09.0 or higher
- CUDA 8.0 or higher
- nvidia-docker
- download CRF++.0.58.tar.gz
put CRF++.0.58.tar.gz invariant2literature/
- http://hgdownload.cse.ucsc.edu/goldenpath/hg19/database/ncbiRefSeq.txt.gz
- http://hgdownload.cse.ucsc.edu/goldenpath/hg19/database/ensGene.txt.gz
- http://hgdownload.cse.ucsc.edu/goldenpath/hg19/database/knownGene.txt.gz
- http://hgdownload.cse.ucsc.edu/goldenpath/hg19/database/kgAlias.txt.gz
- http://hgdownload.cse.ucsc.edu/goldenpath/hg19/database/ensemblToGeneName.txt.gz
- http://hgdownload.cse.ucsc.edu/goldenpath/hg19/database/snp150.txt.gz
- ftp://ftp.ncbi.nih.gov/gene/DATA/GENE_INFO/Mammalia/Homo_sapiens.gene_info.gz
- ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/gene2refseq.gz
- download from ucsc and convert it to fasta format or download from GATK bundle and decompress it.
rename it toucsc.hg19.fasta
(if the filename is notucsc.hg19.fasta
).
- download tmVar 2.0
copytmVarJava/CRF/MentionExtractionUB.Model
tovariant2literature/models/
- download GNormPlus
copyGNormPlusJava/Dictionary/GNR.Model
tovariant2literature/models/
and
copyGNormPlusJava/Dictionary/PT_CTDGene.txt
tovariant2literature/models/
- build docker image by
make build
- compile fasterRCNN by
make compile
- start docker container by
make run
- start mysql docker container by
make run-db
- load data into database by
make load-db
(run only once unless MYSQL_VOLUME is changed)
- put paper directories in input/
- run
make index
- query by
make query
ormake query OUTPUT_FILE=output.txt
- run
make truncate
- run
make rm
- run
make rm-db
This project is licensed under the GPLv3 License - see the LICENSE file for details.
The fasterRCNN implementation used here is written by Jianwei Yang and Jiasen Lu.