del - Githubissues

Phelimb commented 9 years ago

/dev/git/Mykrobe-predictor/bin/Mykrobe.predictor.staph --file 2d.fq.9hours.fastq --install_dir ~/dev/git/Mykrobe-predictor ... "called_variants" :{ "gyrA_S84L" :{ "R_per_cov": "93", "S_per_cov": "0", "R_median_cov": "2", "S_median_cov": "0", "conf": "12", "induced_resistance": "Ciprofloxacin" }

and this is what we get when we subsample to 8 hours

    "called_variants" :{
            "grlA_S80F" :{
                    "R_per_cov": "100",
                    "S_per_cov": "0",
                    "R_median_cov": "2",
                    "S_median_cov": "0",
                    "conf": "12",
            "induced_resistance": "Ciprofloxacin"
            }

iqbal-lab commented 9 years ago

Our mental model of the implementation says that if you detect an allele at covg x, you should always detect it for covg >x. You may or may not call the sample as minor/major resistant, but you should see and report the allele. So there is a disjunct between our mental model and empirical behaviour,

On 21 April 2015 at 10:24, Phelim Bradley notifications@github.com wrote:

/dev/git/Mykrobe-predictor/bin/Mykrobe.predictor.staph --file 2d.fq.9hours.fastq --install_dir ~/dev/git/Mykrobe-predictor ... "called_variants" :{ "gyrA_S84L" :{ "R_per_cov": "93", "S_per_cov": "0", "R_median_cov": "2", "S_median_cov": "0", "conf": "12", "induced_resistance": "Ciprofloxacin" }

and this is what we get when we subsample to 8 hours
"called_variants" :{
        "grlA_S80F" :{
                "R_per_cov": "100",
                "S_per_cov": "0",
                "R_median_cov": "2",
                "S_median_cov": "0",
                "conf": "12",
        "induced_resistance": "Ciprofloxacin"
        }
— Reply to this email directly or view it on GitHub https://github.com/iqbal-lab/Mykrobe-predictor/issues/63.

iqbal-lab commented 9 years ago

Like all the best bugs!

On 21 April 2015 at 10:43, Zamin Iqbal zam.iqbal.lab@gmail.com wrote:

Our mental model of the implementation says that if you detect an allele at covg x, you should always detect it for covg >x. You may or may not call the sample as minor/major resistant, but you should see and report the allele. So there is a disjunct between our mental model and empirical behaviour,

On 21 April 2015 at 10:24, Phelim Bradley notifications@github.com wrote:
/dev/git/Mykrobe-predictor/bin/Mykrobe.predictor.staph --file 2d.fq.9hours.fastq --install_dir ~/dev/git/Mykrobe-predictor ... "called_variants" :{ "gyrA_S84L" :{ "R_per_cov": "93", "S_per_cov": "0", "R_median_cov": "2", "S_median_cov": "0", "conf": "12", "induced_resistance": "Ciprofloxacin" }

and this is what we get when we subsample to 8 hours
"called_variants" :{
        "grlA_S80F" :{
                "R_per_cov": "100",
                "S_per_cov": "0",
                "R_median_cov": "2",
                "S_median_cov": "0",
                "conf": "12",
        "induced_resistance": "Ciprofloxacin"
        }
— Reply to this email directly or view it on GitHub https://github.com/iqbal-lab/Mykrobe-predictor/issues/63.

iqbal-lab / Mykrobe-predictor

del #63