Thanks for your reply!
In fact, I tried to use software ldsc to evaluate the heritability using the LD score computed from ldsc and GCTA respectively and get very different estimates.
Here is the script I use:
# ldsc
python "$src_dir"/ldsc.py \
--h2 $sumstats_dir/cau_ratio_"$cau_ratio"_hsq_"$hsq"_sim_"$sim_i".txt \
--ref-ld-chr $ldsc_ldscore_dir/ \
--w-ld-chr $ldsc_ldscore_dir/ \
--out "$ldsc_out_dir"/cau_ratio_"$cau_ratio"_hsq_"$hsq"_sim_"$sim_i"
# gcta
python "$src_dir"/ldsc.py \
--h2 $sumstats_dir/cau_ratio_"$cau_ratio"_hsq_"$hsq"_sim_"$sim_i".txt \
--ref-ld-chr $gcta_ldscore_dir/ \
--w-ld-chr $gcta_ldscore_dir/ \
--out "$gcta_out_dir"/cau_ratio_"$cau_ratio"_hsq_"$hsq"_sim_"$sim_i"
For ldsc, the output log is like
Total Observed scale h2: 0.6521 (0.0229)
Lambda GC: 3.8227
Mean Chi^2: 5.0959
Intercept: 1.2839 (0.0241)
Ratio: 0.0693 (0.0059)
For gcta, the output log is like
Total Observed scale h2: 1.084 (0.0476)
Lambda GC: 3.8227
Mean Chi^2: 5.0959
Intercept: 1.4536 (0.0326)
Ratio: 0.1107 (0.008)
I used the same data for comparison.

Also, I tried to plot the correlation between the ld scores computed from ldsc and gcta.
在 2018年8月6日星期一 UTC-7下午3:41:08,Raymond Walters写道: