Build a Methylation Hash Table

To run DRAGEN methylation pipeline, you need to build a methylation-specific hash table. Both multi-pass and single-pass options require methylation hash tables. The type of hash table must match the method used during alignment (methylation-mapping-implementation: single-pass| multi-pass). For more information on multi-pass and single-pass methods, see Mapping Method Options.

If using the single-pass method, make sure that your hash table meets the following requirements:

•

Set methylation-mapping-implementation: single-pass.

•

Set ht-methylated-combined=true to generate a combined hash table in the methyl_converted subdirectory.

•

Single-pass mode supports alt-contig references. Each contig in the original FASTA appears twice in the combined reference genome files, one time with each conversion type.

If using the multi-pass method, make sure that your hash table meets the following requirements:

•

Set methylation-mapping-implementation: multi-pass.

•

Set ht-methylated=true to generate the CT_converted and GA_converted subdirectories.

•

Multi-pass mode does not support alt-contig references. The subdirectories contain the genome index for C to T and G to A conversion.

You can set ht-methylated-combined=true and ht-methylated=true together to create a hash table directory that can be used for both methylation-mapping methods.

The following is an example command line for a single-pass hash table.

The --ht-seed-len 27 option is important for optimal results. The --ht-num-threads option enables multiple numbers of threads for faster building.