Different output on example run when sample size is increased by 1
liangyy opened this issue · 2 comments
Hi Natsuhiko,
I am getting familiar with RASQUAL right now and testing the example run. If I increase the sample size from 24 to 25, the top SNP changes. Does it have something to do with setting prior or some internal parameters? So that I should always set the exact number of samples rather than giving rasqual
a rough estimate of the minimum like -m
and -l
?
$ tabix data/chr11.gz 11:2315000-2340000 | rasqual -y data/Y.bin -k data/K.bin -n 24 -j 1 -l 378 -m 600 -s 2316875,2320655,2321750,2321914,2324112 -e 2319151,2320937,2321843,2323290,2324279 --fix-genotype -t
1 rs2521269 11 2321095 C A 0.604167 0.041818 0.965100 -13.4603461799 66.1521580593 0.092827 0.000033 0.525289 2.934533 4.180899 11 83 7 4 2321095 -197.774783 0 0.995314 0.993467
$ tabix data/chr11.gz 11:2315000-2340000 | rasqual -y data/Y.bin -k data/K.bin -n 25 -j 1 -l 378 -m 600 -s 2316875,2320655,2321750,2321914,2324112 -e 2319151,2320937,2321843,2323290,2324279 --fix-genotype -t
1 rs12224967 11 2321284 G A 0.340000 0.978724 0.978100 -11.0378792353 55.5774037017 0.887923 0.000037 0.494012 1.963603 4.263620 11 102 7 4 2321284 -201.188234 0 0.995295 0.992414
If I further increase the sample size, I got an error or some output like
$ tabix data/chr11.gz 11:2315000-2340000 | rasqual -y data/Y.bin -k data/K.bin -n 100 -j 1 -l 378 -m 600 -s 2316875,2320655,2321750,2321914,2324112 -e 2319151,2320937,2321843,2323290,2324279 --fix-genotype -t
gsl: gamma.c:1180: ERROR: error
Default GSL error handler invoked.
Aborted
$ tabix data/chr11.gz 11:2315000-2340000 | rasqual -y data/Y.bin -k data/K.bin -n 200 -j 1 -l 378 -m 600 -s 2316875,2320655,2321750,2321914,2324112 -e 2319151,2320937,2321843,2323290,2324279 --fix-genotype -t
1 SKIPPED 11 -1 N N -1.0 -1.0 -1.0 0.0 0.0 -1.0 -1.0 -1.0 -1.0 -1.0 0 0 -1 -1 -1 0.0 0 -1.0 -1.0
Are they expected?
Thanks!
Yanyu
Hi Yanyu,
I don't imagine any situation in which you only know the lower bound of the sample size for QTL mapping. We know that the test data is composed of the 24 samples, because we collected the 24 samples.
Best regards,
Natsuhiko
Hi Natsuhiko,
Thank you for the explanation! Got your point. I will close this issue.
Yanyu