Weird ALT call not being treated as ALT
mbhall88 opened this issue · 0 comments
mbhall88 commented
I have been looking at the mutation genotype concordance between Illumina and Nanopore and have stumbled across a weird case.
The mutation katG_G1903GCT-C2154209AGC
is called as ALT (1/1) in Illumina, but REF (0/0) on Nanopore. However, the ALT call by mykrobe-Illumina does not appear in the susceptibility information for (isoniazid) resistance...
The command to generate the Illumina file was
$ mykrobe predict -e 0.001 -o /hps/nobackup/iqbal/mbhall/tech_wars/analysis/resistance_prediction/results/mykrobe/predict/illumina/south_africa/R25048/R25048.mykrobe.json -i /hps/nobackup/iqbal/mbhall/tech_wars/data/QC/subsampled/south_africa/illumina/R25048/R25048.subsampled.R1.fastq.gz /hps/nobackup/iqbal/mbhall/tech_wars/data/QC/subsampled/south_africa/illumina/R25048/R25048.subsampled.R2.fastq.gz --force -A --debug --format json --min_proportion_expected_depth 0.20 --sample R25048 --species tb -t 4 -m 4096MB > /hps/nobackup/iqbal/mbhall/tech_wars/analysis/resistance_prediction/logs/rules/mykrobe/illumina/south_africa/R25048.log 2>&1
Here is the entry from Illumina
"katG_G1903GCT-C2154209AGC": {
"variant": "ref-G1903GCT?var_name=C2154209AGC&num_alts=1&ref=NC_000962.3&enum=0&gene=katG&mut=G1903GCT",
"genotype": [
1,
1
],
"genotype_likelihoods": [
-2855.8461119184817,
-99999999,
-1465.7537363958886
],
"info": {
"coverage": {
"reference": {
"percent_coverage": 100.0,
"median_depth": 5,
"min_non_zero_depth": 5,
"kmer_count": 108,
"klen": 21
},
"alternate": {
"percent_coverage": 14.29,
"median_depth": 0,
"min_non_zero_depth": 1,
"kmer_count": 393,
"klen": 22
}
},
"expected_depths": [
24
],
"contamination_depths": [],
"filter": [],
"conf": 1390
},
"_cls": "Call.VariantCall"
},
and for Nanopore
command used
$ mykrobe predict -e 0.08 --ploidy haploid -o /hps/nobackup/iqbal/mbhall/tech_wars/analysis/resistance_prediction/results/mykrobe/predict/nanopore/south_africa/R25048/R25048.mykrobe.json -i /hps/nobackup/iqbal/mbhall/tech_wars/data/QC/subsampled/south_africa/nanopore/R25048/R25048.subsampled.fastq.gz --force -A --debug --format json --min_proportion_expected_depth 0.20 --sample R25048 --species tb -t 4 -m 4096MB > /hps/nobackup/iqbal/mbhall/tech_wars/analysis/resistance_prediction/logs/rules/mykrobe/nanopore/south_africa/R25048.log 2>&1
"katG_G1903GCT-C2154209AGC": {
"variant": "ref-G1903GCT?var_name=C2154209AGC&num_alts=1&ref=NC_000962.3&enum=0&gene=katG&mut=G1903GCT",
"genotype": [
0,
0
],
"genotype_likelihoods": [
-157.54616831189273,
-10013.28801821664
],
"info": {
"coverage": {
"reference": {
"percent_coverage": 100.0,
"median_depth": 96,
"min_non_zero_depth": 84,
"kmer_count": 1895,
"klen": 21
},
"alternate": {
"percent_coverage": 9.52,
"median_depth": 0,
"min_non_zero_depth": 87,
"kmer_count": 174,
"klen": 22
}
},
"expected_depths": [
100.0
],
"contamination_depths": [],
"filter": [],
"conf": 9856
},
"_cls": "Call.VariantCall"
},
I guess the two major questions here are
- Why was this variant not added as evidence for isoniazid resistance in Illumina?
- Why was this variant called homozygous ALT?