Help!Algorithm failed to converge after trying different initial values

[jy-best]

Hi,
I am using your tool to calculate my own summary data from UKB. Because my raw data is the imputed genotype data, i used the UKB_imputed_SVD_eigen99_extraction as panel.
I have calculate about two hundred pairs of diseases and it works well.
But three summary statistics still have error, which is the same format of other summary statistics. Like this:
Analysis starts on Tue Mar 23 16:30:24 2021
1014352 out of 1029876 (98.49%) SNPs in reference panel are available in GWAS 1.
1017202 out of 1029876 (98.77%) SNPs in reference panel are available in GWAS 2.
Warning: More than 1% SNPs in reference panel are missed in GWAS 1. This may generate bias in estimation. Please make sure that you are using correct reference panel.
Warning: More than 1% SNPs in reference panel are missed in GWAS 2. This may generate bias in estimation. Please make sure that you are using correct reference panel.
Estimation is ongoing ... 100%

Integrating piecewise results
Error in HDL.rg(summary1, summary2, :
Algorithm failed to converge after trying different initial values.

Although someone has raise this question similarly, but i checked carefully and couldn't find any other wrong operation.
Looking forward to your reply!

[yurimilaneschi]

I had the same issue with two highly correlated traits. when rerunning the rg with LDSC i had indeed an estimate statistically not different form 1.
consistently, you get the same error if you run hdl for the same trait (non convergence for rg not different form 1) . may be this the issue?
any help from the developer would be highly appreciated, thank you for this great tool.

[anbai106]

Same here #6