# Create new version HLA reference for Kourami
###### tags: `c4lab`

Kourami
https://github.com/Kingsford-Group/kourami
Lee, H., & Kingsford, C. Kourami: graph-guided assembly for novel human leukocyte antigen allele discovery. Genome Biology 19(16), 2018

I follow this tutorial https://github.com/Kingsford-Group/kourami/blob/master/preprocessing.md#2-creating-kourami-hla-panel-and-merged-msas-from-another-version-release-of-imgthla-db-or-a-custom-version

## Prerequested

You can successful run Kourmai for HLA typing.

i.e. The below command should work

java -jar kourami-0.9.6/build/Kourami.jar \
    -d ${kourami_index}
    ${sample}_on_KouramiPanel.bam
    -o ${sample}.kourami

Where `kourami-0.9.6/build/Kourami.jar` are downloaded from https://github.com/Kingsford-Group/kourami/releases/tag/v0.9.6

`${sample}_on_KouramiPanel.bam` is created followed by https://github.com/Kingsford-Group/kourami/blob/master/preprocessing.md

`kourami_index` is the index path for Kourami. Default is `kourami/db` if you use `download_panel.sh`.

## Download HLA
### Latest version

You can download from IMGT/HLA FTP
http://ftp.ebi.ac.uk/pub/databases/ipd/imgt/hla/

Find the alignment data `Alignments_Rel_3450.zip` in latest version (e.g. 3.45.0 at 2021/08/12)

And `wmda/hla_nom_g.txt` (This file should be same version of Alignments data)

wget ftp://ftp.ebi.ac.uk/pub/databases/ipd/imgt/hla/Alignments_Rel_3450.zip
unzip Alignments_Rel_3450.zip
wget ftp://ftp.ebi.ac.uk/pub/databases/ipd/imgt/hla/wmda/hla_nom_g.txt

### Previous version

You can find older version in github repo https://github.com/ANHIG/IMGTHLA by searching in branch.

In this exmpale, I want to use `3.42.0`, then choose `3420` branch(https://github.com/ANHIG/IMGTHLA/tree/3420), then find and download the alignment data and hla_nom_g

wget https://github.com/ANHIG/IMGTHLA/raw/3420/Alignments_Rel_3420.zip
unzip Alignments_Rel_3420.zip
wget https://github.com/ANHIG/IMGTHLA/raw/3420/wmda/hla_nom_g.txt

## Build Index

(I use 3.42.0 as example)

java -cp kourami-0.9.6/build/Kourami.jar FormatIMGT Alignments_Rel_3420 v3.42.0 tmp
mv tmp/v3.42.0 mydb
cat mydb/*.merged.fa resources/HLA_decoys.fa | gzip > mydb/All_FINAL_with_Decoy.fa.gz
cp hla_nom_g.txt mydb
bwa index mydb/All_FINAL_with_Decoy.fa.gz

If kourami throw this error, it's a [known bug](https://github.com/Kingsford-Group/kourami/issues/19)

REF SEQ names differs :
(nuc):Y*01:01
(gen):Y
java.lang.NullPointerException: Cannot invoke "LogHandler.appendln(String)" because "HLA.log" is null
        at Sequence.processBlock(Sequence.java:554)
        at Sequence.<init>(Sequence.java:498)
        at MergeMSFs.mergeAndAdd(MergeMSFs.java:383)
        at MergeMSFs.mergeAndAdd(MergeMSFs.java:372)  
        at MergeMSFs.merge(MergeMSFs.java:298)
        at FormatIMGT.processGene(FormatIMGT.java:199)
        at FormatIMGT.main(FormatIMGT.java:100)         

You can modified `Alignments_Rel_3420/Y_gen.txt` to remove leading deletion **if Y is NOT your target HLA**.

 gDNA              0   
                   |   
 Y*01:01           | ATGGCGGTC 
 Y*02:01           | ---------
 Y*03:01           | --------- 

## Try it

kourami_index='mydb'
java -jar kourami-0.9.6/build/Kourami.jar \
    -d ${kourami_index}
    ${sample}_on_KouramiPanel.bam
    -o ${sample}.kourami