Credible sets for Finemap-inf

[sayangsep]

Hi, I was trying to run Finemap-inf and generate credible sets. But, I cannot find the outputs for credible sets. Could you please guide me on this?

[cuiran]

Hi, we currently only have credible set information for SuSiE-inf not FINEMAP-inf. If you run the run_fine_mapping.py script, The SuSiE-inf credible set information can be found in the column named ‘cs’. The different numerical values indicate which credible set the corresponding SNP is in, -1 value means the SNP is not in any credible set. If you run susieinf.py directly, the credible set information is stored in the key ‘cred’ in the outputted dictionary. Let me know if that helps.

[sayangsep]

Thank you for your reply! I have another question: While running susie-inf is there a way to fix $s^2$ (i.e. variance of fixed effects) and $\tau^2$ (variance of residual noise)?

[cuiran]

Yes, you can specify ssq and tausq in the function susie in susieinf.py https://github.com/FinucaneLab/fine-mapping-inf/blob/4af3e6ebf04f372bcb641977601a2b65a4080237/susieinf/susieinf.py#L79

[sayangsep]

Thanks for your reply! While running susie-inf I am encountering the following error. Can you please tell me what's causing this error? I am also attaching my z-file and LD file for your reference.

################## Error ################## Reading summary statistics from file /mnt/sdb1/sayan/finemapping_mixed/data/data21/setting60/Simulation_data1.z 0 1 0 rs0 -13.799580 1 rs1 16.040077 2 rs2 -8.640558 3 rs3 4.779611 4 rs4 36.948448 .. ... ... 995 rs995 -7.753660 996 rs996 1.300606 997 rs997 -7.236649 998 rs998 -2.895833 999 rs999 1.436788

[1000 rows x 2 columns] 1000 SNPs in summary statistics file Reading in LD matrix from file /mnt/sdb1/sayan/finemapping_mixed/data/data21/setting60/Simulation_data1.ld Reading in LD matrix took 0.12 seconds Performing eigen decomposition Eigen decomposition took 0.18 seconds Running susieinf

---

• SuSiE-inf
• Version 1.2
• (C) Zhou Fan

  ---

  /home/sayan/anaconda3/envs/finemapping_inf/lib/python3.9/site-packages/susieinf.py:170: RuntimeWarning: invalid value encountered in log return -scipy.special.logsumexp(-0.5np.log(1+xdiagXtOmegaX ) \ /home/sayan/anaconda3/envs/finemapping_inf/lib/python3.9/site-packages/susieinf.py:181: RuntimeWarning: invalid value encountered in log logPIP = XtOmegar*2/(2omega[:,l])-0.5np.log(omega[:,l]ssq[l] )+logpi0 Iteration 0 Update s^2 for effect 0 to 0.999995 Update s^2 for effect 1 to 0.023139 Update s^2 for effect 2 to 0.008940 Update s^2 for effect 3 to 0.014246 Update s^2 for effect 4 to 0.013241 Update s^2 for effect 5 to 0.004850 Update s^2 for effect 6 to 0.007866 Update s^2 for effect 7 to 0.002449 Update s^2 for effect 8 to 0.003442 Update s^2 for effect 9 to 0.002888 Update s^2 for effect 10 to 0.000793 Update s^2 for effect 11 to 0.000363 Update s^2 for effect 12 to 0.000238 Update s^2 for effect 13 to 0.000150 Update s^2 for effect 14 to 0.000077 Update s^2 for effect 15 to 0.000026 Update s^2 for effect 16 to 0.000008 Update s^2 for effect 17 to 0.000006 Update s^2 for effect 18 to 0.000006 Update s^2 for effect 19 to 0.000006 Update (sigma^2,tau^2) to (-0.113639,0.000000e+00) Maximum change in PIP: 0.999000 Iteration 1 WARNING: s^2 update for iteration 1, effect 0 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 1 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 2 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 3 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 4 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 5 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 6 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 7 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 8 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 9 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 10 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 11 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 12 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 13 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 14 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 15 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 16 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 17 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 18 failed to converge; keeping previous parameters WARNING: s^2 update for iteration 1, effect 19 failed to converge; keeping previous parameters Traceback (most recent call last): File "/home/sayan/hdd/sayan/finemapping_mixed/plos_experiments/susie_inf/run_fine_mapping.py", line 199, in susie_output = susie(z, args.meansq, args.n, args.num_sparse_effects, LD=None, V=V, Dsq=Dsq, File "/home/sayan/anaconda3/envs/finemapping_inf/lib/python3.9/site-packages/susieinf.py", line 186, in susie (sigmasq,tausq) = MoM(PIP,mu,omega,sigmasq,tausq,n,V,Dsq,VtXty,Xty,yty, File "/home/sayan/anaconda3/envs/finemapping_inf/lib/python3.9/site-packages/susieinf.py", line 29, in MoM sol = scipy.linalg.solve(A,x) File "/home/sayan/anaconda3/envs/finemapping_inf/lib/python3.9/site-packages/scipy/linalg/_basic.py", line 144, in solve b1 = atleast_1d(_asarray_validated(b, check_finite=check_finite)) File "/home/sayan/anaconda3/envs/finemapping_inf/lib/python3.9/site-packages/scipy/_lib/_util.py", line 252, in _asarray_validated a = toarray(a) File "/home/sayan/anaconda3/envs/finemapping_inf/lib/python3.9/site-packages/numpy/lib/function_base.py", line 628, in asarray_chkfinite raise ValueError( ValueError: array must not contain infs or NaNs Simulation_data.zip