Hi,
I am trying to use optitype for HLA typing for targeted amplicon sequencing samples applied on ctDNA targeting exonic regions of dna repair related genes.
For this, we usually do UMI resolution to get consensus reads and then align them and use those BAM files for HLA typing. Visualising these BAMs on IGV shows large coverage on IGV. However, when I try to use optitype for HLA typing, optitype ends up dropping most of the reads and is doing HLA calling from as low as 1 read at times but mostly around 10 to 20.
Mapping with RazerS3 using 4 threads...
0:00:00.38 Mapping 2590368-KCL732-T.1.fq to GEN reference...
0:00:01.99 Mapping 2590368-KCL732-T.2.fq to GEN reference...
0:00:03.64 Generating binary hit matrix.
0:00:03.67 Loading /data/scratch/DBC/UBCN/BCRBIOIN/SHARED/analysis/ChrisLord/tas_reversion_ctDNA_panel/data/RCBH_7779_AG/hla_typing//2590368-KCL732-T/2590368-KCL732-T_1.bam started. Number of HLA reads loaded (updated every thousand):
0:00:03.67 1 reads loaded. Creating dataframe...
0:00:03.83 Dataframes created. Shape: 1 x 11179, hits: 999 (1019), sparsity: 1 in 10.97
0:00:03.86 Loading /data/scratch/DBC/UBCN/BCRBIOIN/SHARED/analysis/ChrisLord/tas_reversion_ctDNA_panel/data/RCBH_7779_AG/hla_typing//2590368-KCL732-T/2590368-KCL732-T_2.bam started. Number of HLA reads loaded (updated every thousand):
0:00:03.86 1 reads loaded. Creating dataframe...
0:00:03.97 Dataframes created. Shape: 1 x 11179, hits: 655 (655), sparsity: 1 in 17.07
0:00:03.98 Alignment pairing completed. 1 paired, 0 unpaired, 0 discordant
0:00:04.96 Temporary pruning of identical rows and columns
0:00:04.96 Size of mtx with unique rows and columns: (1, 2)
0:00:04.96 Determining minimal set of non-overshadowed alleles
0:00:04.97 Keeping only the minimal number of required alleles (1,)
0:00:04.97 Creating compact model...
Starting ILP solver with 1 threads...
0:00:04.97 Initializing OptiType model...
0:00:04.98 Result dataframe has been constructed...
However, when I visualise on IGV, there are lots of reads that are aligned to the HLA regions.
I was wondering is optitype not optimised for samples from targeted sequencing?
Hi,
I am trying to use optitype for HLA typing for targeted amplicon sequencing samples applied on ctDNA targeting exonic regions of dna repair related genes.
For this, we usually do UMI resolution to get consensus reads and then align them and use those BAM files for HLA typing. Visualising these BAMs on IGV shows large coverage on IGV. However, when I try to use optitype for HLA typing, optitype ends up dropping most of the reads and is doing HLA calling from as low as 1 read at times but mostly around 10 to 20.
Mapping with RazerS3 using 4 threads...
0:00:00.38 Mapping 2590368-KCL732-T.1.fq to GEN reference...
0:00:01.99 Mapping 2590368-KCL732-T.2.fq to GEN reference...
0:00:03.64 Generating binary hit matrix.
0:00:03.67 Loading /data/scratch/DBC/UBCN/BCRBIOIN/SHARED/analysis/ChrisLord/tas_reversion_ctDNA_panel/data/RCBH_7779_AG/hla_typing//2590368-KCL732-T/2590368-KCL732-T_1.bam started. Number of HLA reads loaded (updated every thousand):
0:00:03.67 1 reads loaded. Creating dataframe...
0:00:03.83 Dataframes created. Shape: 1 x 11179, hits: 999 (1019), sparsity: 1 in 10.97
0:00:03.86 Loading /data/scratch/DBC/UBCN/BCRBIOIN/SHARED/analysis/ChrisLord/tas_reversion_ctDNA_panel/data/RCBH_7779_AG/hla_typing//2590368-KCL732-T/2590368-KCL732-T_2.bam started. Number of HLA reads loaded (updated every thousand):
0:00:03.86 1 reads loaded. Creating dataframe...
0:00:03.97 Dataframes created. Shape: 1 x 11179, hits: 655 (655), sparsity: 1 in 17.07
0:00:03.98 Alignment pairing completed. 1 paired, 0 unpaired, 0 discordant
0:00:04.96 Temporary pruning of identical rows and columns
0:00:04.96 Size of mtx with unique rows and columns: (1, 2)
0:00:04.96 Determining minimal set of non-overshadowed alleles
0:00:04.97 Keeping only the minimal number of required alleles (1,)
0:00:04.97 Creating compact model...
Starting ILP solver with 1 threads...
0:00:04.97 Initializing OptiType model...
0:00:04.98 Result dataframe has been constructed...
However, when I visualise on IGV, there are lots of reads that are aligned to the HLA regions.
I was wondering is optitype not optimised for samples from targeted sequencing?