Version : v4.1.0 Describe the bug

Here are the parameters: <div class="snippet-clipboard-content notranslate positio

<a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/us

Zero matches between base and comp about truvari HOT 4 CLOSED

jiadong324 commented on June 21, 2024

Zero matches between base and comp

from truvari.

Comments (4)

ACEnglish commented on June 21, 2024

Could you let me know what parameters you used? They can be found at the head of the bench_result/log.txt. Something like

2024-02-07 15:45:59,529 [INFO] Params:
{
    "base": "benchmark.HG002.GRCh38.norm.vcf.gz",
    "comp": "HG002.GRCh38.vcf.gz",
    "output": "bench_result",
    "includebed": "GIAB-Q100.GRCh38.bed",
    "extend": 0,
    "debug": false,
    "reference": null,
    "refdist": 500,
    "pctseq": 0.7,
    "minhaplen": 50,
    "pctsize": 0.7,
    "pctovl": 0.0,
    "typeignore": true,
    "chunksize": 1000,
    "bSample": "syndip",
    "cSample": "HG002",
    "dup_to_ins": true,
    "sizemin": 1,
    "sizefilt": 1,
    "sizemax": 1000,
    "passonly": false,
    "no_ref": "a",
    "pick": "single",
    "check_monref": true,
    "check_multi": true
}

from truvari.

jiadong324 commented on June 21, 2024

Here are the parameters:

{
    "base": "assembly.svs.vcf.gz",
    "comp": "nanovar.vcf.gz",
    "output": "truvari_bench/nanovar_hifi",
    "includebed": null,
    "extend": 0,
    "debug": false,
    "reference": null,
    "refdist": 500,
    "pctseq": 0.7,
    "minhaplen": 50,
    "pctsize": 0.7,
    "pctovl": 0.0,
    "typeignore": false,
    "chunksize": 1000,
    "bSample": "HG00171",
    "cSample": "HG00171.mm2.sorted",
    "dup_to_ins": false,
    "sizemin": 50,
    "sizefilt": 30,
    "sizemax": 50000,
    "passonly": true,
    "no_ref": false,
    "pick": "single",
    "check_monref": true,
    "check_multi": true
}

from truvari.

ACEnglish commented on June 21, 2024

Hi,

So the "pctseq": 0.7, is the problem here. Truvari needs sequence resolved variants to calculate sequence similarity. If sequence similarity is on, then symbolic alts are filtered out (there should be a warning in the log file). If you turn off sequence similarity with truvari bench --pctseq 0, then the symbolic alts are not removed.

Have a great day,
~/Adam English

from truvari.

jiadong324 commented on June 21, 2024

@ACEnglish

It is fixed when I use --pctseq 0. Thanks!

from truvari.

Recommend Projects