Post by jverlouw on Nov 3, 2017 13:58:53 GMT
Dear Jian,
Currently we are trying to use your GREML-LDMS to perform a heritability analysis on our HRC1.1 imputed data. Unfortunately we hit a bit of a problem right from the start, namely the --ld-score-region command not producing any output, without parsing any errors. Do you have any idea what might cause this behavior? Below are the command used and a tail of the strace.
Command:
Alternatively I tried to perform roughly the same analysis using the normal LD command as follows:
However, this produces a segmentation fault right after finishing loading in the dosage data (without threading).
Might this be caused by us using mach dosage files instead of plink format dosage/genotype files?
Thanks in forward for looking into this!
Currently we are trying to use your GREML-LDMS to perform a heritability analysis on our HRC1.1 imputed data. Unfortunately we hit a bit of a problem right from the start, namely the --ld-score-region command not producing any output, without parsing any errors. Do you have any idea what might cause this behavior? Below are the command used and a tail of the strace.
Command:
gcta64_1902 --dosage-mach-gz Dosages/chr22.mach.dose.gz Dosages/chr22.mach.info --ld-score-region 200 --out chr22
Strace tail:...
read(3, "\324\376h\25\217O(c\247\376\346\336S\264\305\271m\2777z\240\257\270<\\\212\30\362\252\3465<"..., 16384) = 16384
read(3, "\370[8\366\7\377\340\247\300\260|K\214o\207\360\276\333\353~>S\220\2530&y\365s\273;\240"..., 16384) = 1192
read(3, "", 12288) = 0
read(3, "", 16384) = 0
close(3) = 0
munmap(0x7f681666c000, 4096) = 0
write(1, "Imputed dosage data for 3048 ind"..., 93) = 93
brk(0x1857ec000) = 0x1857ec000
brk(0x1858ec000) = 0x1858ec000
brk(0x1859ec000) = 0x1859ec000
brk(0x185aec000) = 0x185aec000
brk(0x185cec000) = 0x185cec000
brk(0x185eec000) = 0x185eec000
brk(0x1860ec000) = 0x1860ec000
brk(0x1862ec000) = 0x1862ec000
brk(0x1866ec000) = 0x1866ec000
brk(0x186aec000) = 0x186aec000
brk(0x186eec000) = 0x186eec000
brk(0x1872ec000) = 0x1872ec000
brk(0x187aec000) = 0x187aec000
brk(0x1882ec000) = 0x1882ec000
brk(0x188aec000) = 0x188aec000
brk(0x1892ec000) = 0x1892ec000
brk(0x18a2ec000) = 0x18a2ec000
brk(0x18b2ec000) = 0x18b2ec000
brk(0x18c2ec000) = 0x18c2ec000
brk(0x18d2ec000) = 0x18d2ec000
mmap(NULL, 33558528, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f680c699000
mmap(NULL, 33558528, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f680a698000
mmap(NULL, 33558528, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f6808697000
mmap(NULL, 33558528, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f6806696000
brk(0x18a32c000) = 0x18a32c000
munmap(0x7f680c699000, 33558528) = 0
munmap(0x7f6806696000, 33558528) = 0
munmap(0x7f6808697000, 33558528) = 0
brk(0x18636c000) = 0x18636c000
munmap(0x7f680a698000, 33558528) = 0
brk(0x7f21000) = 0x7f21000
munmap(0x7f680e69a000, 16789504) = 0
munmap(0x7f681666d000, 16789504) = 0
munmap(0x7f68136a7000, 33558528) = 0
munmap(0x7f680f69d000, 33558528) = 0
munmap(0x7f681169e000, 33558528) = 0
brk(0x3ef5000) = 0x3ef5000
write(1, "\33[0m\33[0m\n", 9) = 9
clock_gettime(CLOCK_REALTIME, {1509716774, 175982631}) = 0
stat("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2923, ...}) = 0
write(1, "\33[0m\33[0mAnalysis finished: Fri N"..., 56) = 56
clock_gettime(CLOCK_MONOTONIC, {23679062, 924468007}) = 0
write(1, "\33[0m\33[0mComputational time: 3 mi"..., 51) = 51
exit_group(0) = ?
+++ exited with 0 +++
Alternatively I tried to perform roughly the same analysis using the normal LD command as follows:
gcta64_1902 --dosage-mach-gz Dosages/chr22.mach.dose.gz Dosages/chr22.mach.info --ld-score --ld-wind 200 --ld-rsq-cutoff 0.01 --out chr22
However, this produces a segmentation fault right after finishing loading in the dosage data (without threading).
Might this be caused by us using mach dosage files instead of plink format dosage/genotype files?
Thanks in forward for looking into this!