quickhla is inspired by https://github.com/ExpressionAnalysis/HLAProfiler
and https://sourceforge.net/projects/targt-pipeline/
to use k-mer content of next generation sequencing reads and subsequent mapping to classify HLA types in a sample.
quickhla heavily relies on the data from the IPD-IMGT/HLA database.
First, the HLA k-mer databases need to be build (can take up to 30 minutes):
# download hla fasta files from ftp://ftp.ebi.ac.uk/pub/databases/ipd/imgt/hla/fasta/
wget ftp://ftp.ebi.ac.uk/pub/databases/ipd/imgt/hla/fasta/hla_gen.fasta
wget ftp://ftp.ebi.ac.uk/pub/databases/ipd/imgt/hla/fasta/hla_nuc.fasta
# build databases with k-mer length of 35
quickhla build -gen hla_gen.fasta -nuc hla_nuc.fasta
# build databases with k-mer length of 35 using 12 threads
quickhla build -gen hla_gen.fasta -nuc hla_nuc.fasta -t 12see help
quickhla build -h
usage: quickhla <sub-script> [options] [<arguments>...] build
[-h] [-gen GEN] [-nuc NUC] [-o O] [-kb KB] [-bb BB] [-hb HB] [-kl KL]
[-ml ML] [-ms MS]
optional arguments:
-h, --help show this help message and exit
-gen GEN input gen file [mandatory]
-nuc NUC input nuc file [mandatory]
-o O output kraken2 database directory [default: hla.db]
-kb KB specify kraken2-build binary [if not given assumes to be in
PATH]
-bb BB specify bowtie2-build binary [if not given assumes to be in
PATH]
-hb HB specify hisat2-build binary [if not given assumes to be in PATH]
-kl KL kraken2 kmer-len array [default: 35] can be an array of values
and needs to be same length as -ml and -ms
-ml ML kraken2 minimizer-len array [default: 31] if array needs to be
same length as -kl and -ms
-ms MS kraken2 minimizer-space array [default: 7] if array needs to be
same length as -kl and -ml
-t T specify number threads [default: 1]
One can also create the consensus sequences for the different digit levels 2d, 4d, 6d and subsequently use these files to build the database.
mkdir msf
cd msf
wget -A msf -l 1 -r -nd --no-parent ftp://ftp.ebi.ac.uk/pub/databases/ipd/imgt/hla/msf/
quickhla msf -i .
cd ..
# 2d
quickhla build -gen msf/gen_2d.consensus.fasta -nuc msf/nuc_2d.consensus.fasta -t 12 -o msf.2d.db
# 4d
quickhla build -gen msf/gen_4d.consensus.fasta -nuc msf/nuc_4d.consensus.fasta -t 12 -o msf.4d.db
# 6d
quickhla build -gen msf/gen_6d.consensus.fasta -nuc msf/nuc_6d.consensus.fasta -t 12 -o msf.6d.db
For paired reads FORWARD.fq and REVERSE.fq using the standard nuc database created for four-digits 4d , the basic usage is:
quickhla classify -f FORWARD.fq -r REVERSE.fq -d hla.db -db hla.nuc.4d.35 -n 4 -t 12see help
quickhla classify -h
usage: quickhla <sub-script> [options] [<arguments>...] classify
[-h] [-f F] [-r R] [-d D] [-db DB] [-o O] [-t T] [-kb KB] [-alg ALG]
[-bb BB] [-bo BO] [-hb HB] [-ho HO] [-n N]
optional arguments:
-h, --help show this help message and exit
-f F specify forward fastq [mandatory]
-r R specify reverse fastq [mandatory]
-d D specify db directory [mandatory]
-db DB specify db for classification [mandatory]
-o O specify output prefix [default: out]
-t T specify number threads [default: 1]
-kb KB specify kraken2 binary [if not given assumes to be in PATH]
-alg ALG specify aligner [default: hisat2]
-w apply weights on read counts [default: False]
-bb BB specify bowtie2 binary [if not given assumes to be in PATH]
-bo BO specify bowtie2 options [default: --very-fast --no-unal
-k 1000]
-hb HB specify hisat2 binary [if not given assumes to be in PATH]
-ho HO specify hisat2 options [default: --fast -k 1000]
-n N specify number of top hits to report [default: show all]
quickhla outputs two plain text files with the HLA classifications (1. from k-mer and 2. from mapping)
The output are:
- k-mer results (number of fragments per class)
Example shows the top 2 hits per HLA class out.hla.nuc.4d.35.report
#quickhla classify -f SRR099966_1.filt.fastq.gz -r SRR099966_2.filt.fastq.gz -d hla.db -db hla.nuc.4d.35 -n 4 -t 12
99.90 99117470 99117470 U 0 unclassified
0.10 98293 10992 R 1 hlaroot
0.01 10824 2831 K 2 A
0.00 2464 196 P 1424 A*02
0.00 356 356 C 1425 A*02:01
0.00 276 276 C 4832 A*02:874
0.00 944 13 P 8499 A*24
0.00 212 212 C 8500 A*24:02
0.00 56 56 C 9843 A*24:371
0.01 10476 3261 K 16304 B
0.00 1155 91 P 29130 B*44
0.00 471 471 C 29320 B*44:03
0.00 75 75 C 30293 B*44:329
0.00 821 9 P 16305 B*07
0.00 133 133 C 16306 B*07:02
0.00 87 87 C 17422 B*07:338
0.01 10549 2352 K 36025 C
0.00 2833 388 P 43684 C*07
0.00 455 455 C 43984 C*07:02
0.00 135 135 C 43685 C*07:01
0.00 1274 70 P 37661 C*03
0.00 254 254 C 39237 C*03:418
0.00 196 196 C 37885 C*03:04
0.01 8596 1785 K 68214 DQB1
0.00 2114 247 P 68874 DQB1*03
0.00 206 206 C 68875 DQB1*03:01
0.00 117 117 C 70083 DQB1*03:369
0.00 1602 297 P 70844 DQB1*05
0.00 205 205 C 70943 DQB1*05:02
0.00 165 165 C 70845 DQB1*05:01
0.01 6789 2058 K 51886 DRB1
0.00 592 14 P 58657 DRB1*15
0.00 85 85 C 58889 DRB1*15:114
0.00 68 68 C 58832 DRB1*15:08
0.00 543 107 P 53018 DRB1*04
0.00 44 44 C 53163 DRB1*04:05
0.00 34 34 C 53892 DRB1*04:286N
- mapping results (number of reads per class)
Example shows the top 4 hits per HLA class and digit level out.hla.nuc.4d.35.class.txt
#quickhla classify -f SRR099966_1.filt.fastq.gz -r SRR099966_2.filt.fastq.gz -d hla.db -db hla.nuc.4d.35 -n 4 -t 12
A 311503
A*01 122814
A*11 77905
A*03 73636
A*30 31114
A*01:01 48306
A*03:01 27812
A*11:01 27718
A*30:02 5378
A*01:01:01 24741
A*03:01:01 16294
A*11:01:01 14110
A*30:02:01 3309
A*01:01:71:NA 488
A*01:251:NA:NA 343
A*01:01:01:24 342
A*01:342:NA:NA 341
B 143234
B*15 27971
B*08 22800
B*57 16677
B*44 9919
B*08:01 7046
B*57:01 5946
B*15:01 4297
B*18:01 3710
B*08:01:01 3665
B*18:01:01 2500
B*57:01:01 2434
B*15:01:01 2236
B*57:03:04:NA 144
B*57:01:01:14 127
B*57:01:01:20 127
B*57:99:NA:NA 127
C 280304
C*07 193392
C*06 33631
C*12 18817
C*05 10718
C*07:01 37102
C*07:02 36435
C*06:02 11190
C*07:04 5048
C*07:01:01 21353
C*07:02:01 20340
C*06:02:01 4978
C*05:01:01 2533
C*07:02:01:17N 230
C*07:777:NA:NA 223
C*07:01:01:47 222
C*07:109:02:NA 222
DQB1 92545
DQB1*03 43277
DQB1*02 27180
DQB1*04 10886
DQB1*06 10320
DQB1*02:01 5047
DQB1*03:02 4872
DQB1*03:03 4636
DQB1*03:01 3725
DQB1*03:02:01 1718
DQB1*03:01:01 1535
DQB1*04:02:01 1409
DQB1*02:02:01 1081
DQB1*02:59:NA:NA 254
DQB1*02:01:01:01 253
DQB1*02:01:01:04 250
DQB1*02:01:01:02 249
DRB1 272859
DRB1*13 48947
DRB1*11 47441
DRB1*03 36771
DRB1*14 36447
DRB1*07:01 8146
DRB1*03:01 7646
DRB1*11:01 7551
DRB1*13:01 7468
DRB1*14:54:01 4685
DRB1*13:02:01 4060
DRB1*13:01:01 3660
DRB1*09:01:02 3330
DRB1*09:01:02:05 668
DRB1*07:01:01:04 667
DRB1*07:01:01:05 667
DRB1*09:01:02:02 667
quickhla requires:
- kraken2 https://ccb.jhu.edu/software/kraken2/
- bowtie2 http://bowtie-bio.sourceforge.net/bowtie2/index.shtml
- hisat2 https://daehwankimlab.github.io/hisat2/
- biopython https://biopython.org/
- numpy https://numpy.org/
You can install them manually (see individual download pages):
- kraken2 https://github.com/DerrickWood/kraken2
- bowtie2 https://sourceforge.net/projects/bowtie-bio/files/bowtie2/2.4.2
- hisat2 https://daehwankimlab.github.io/hisat2/download/
- biopython https://biopython.org/wiki/Download
- numpy https://numpy.org/install/
Or you can install them via conda:
- kraken2 https://anaconda.org/bioconda/kraken2
- bowtie2 https://anaconda.org/bioconda/bowtie2
- hista2 https://anaconda.org/biobuilds/hisat2
- biopython https://anaconda.org/anaconda/biopython
- numpy https://anaconda.org/anaconda/numpy
conda install -c bioconda kraken2
conda install -c bioconda bowtie2
conda install -c biobuilds hisat2
conda install -c anaconda biopython
conda install -c anaconda numpy
git clone https://gitlab.gwdg.de/kristian.ullrich/quickhla.git
cd quickhla
python setup.py installconda create -n quickhla python=3.8
conda activate quickhla
conda install -c bioconda kraken2
conda install -c bioconda bowtie2
conda install -c biobuilds hisat2
conda install -c anaconda biopython
conda install -c anaconda numpy
git clone https://gitlab.gwdg.de/kristian.ullrich/quickhla.git
cd quickhla
python setup.py installTest HLA typing for individuals from https://www.internationalgenome.org/
# download test data
wget ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/phase3/data/HG00100/sequence_read/SRR099966_1.filt.fastq.gz
wget ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/phase3/data/HG00100/sequence_read/SRR099966_2.filt.fastq.gz
# classify reads - top 4 hits
quickhla classify -f SRR099966_1.filt.fastq.gz -r SRR099966_2.filt.fastq.gz -d hla.db -db hla.nuc.4d.35 -n 4 -t 12
# classify reads - top 4 hits - read counts weighted
quickhla classify -f SRR099966_1.filt.fastq.gz -r SRR099966_2.filt.fastq.gz -d hla.db -db hla.nuc.4d.35 -n 4 -t 12 -w
If you would like to contribute to CRBHits, please file an issue so that one can establish a statement of need, avoid redundant work, and track progress on your contribution.
Before you do a pull request, you should always file an issue and make sure that someone from the CRBHits developer team agrees that it’s a problem, and is happy with your basic proposal for fixing it.
Once an issue has been filed and we've identified how to best orient your contribution with package development as a whole, fork the main repo, branch off a feature branch from master, commit and push your changes to your fork and submit a pull request for quickhla:master.
By contributing to this project, you agree to abide by the Code of Conduct terms.
Please report any errors or requests regarding quickhla to Kristian Ullrich (ullrich@evolbio.mpg.de)
or use the issue tracker at https://gitlab.gwdg.de/kristian.ullrich/quickhla/issues
This repository adhere to Contributor Covenant code of conduct for in any interactions you have within this project. (see Code of Conduct)
See also the policy against sexualized discrimination, harassment and violence for the Max Planck Society Code-of-Conduct.
By contributing to this project, you agree to abide by its terms.
-
Buchkovich M.L., Brown C.C., Robasky K., Chai S., Westfall S., Vincent B.G., Weimer E.T., Powers J.G. (2017) HLAProfiler utilizes k-mer profiles to improve HLA calling accuracy for rare and common alleles in RNA-seq data. Genome Med, 9, 86. https://doi.org/10.1186/s13073-017-0473-6
-
Pierini F., Nutsua M., Böhme L., Özer O., Bonczarowska J., Susat J., Franke A., Nebel A., Krause-Kyora B., Lenz T.L. (2020) Targeted analysis of polymorphic loci from low-coverage shotgun sequence data allows accurate genotyping of HLA genes in historical human populations. Sci Rep, 10, 7339. https://doi.org/10.1038/s41598-020-64312-w
-
Robinson J., Barker D.J., Georgiou X., Cooper M.A., Flicek P., Marsh S.G.E. (2020) IPD-IMGT/HLA Database Nucleic Acids Research, 48(D1), D948–D955. https://doi.org/10.1093/nar/gkz950