+-----------------------------+ | ___ | | BOLT-LMM, v2.3.1 /_ / | | December 19, 2017 /_/ | | Po-Ru Loh // | | / | +-----------------------------+ Copyright (C) 2014-2017 Harvard University. Distributed under the GNU GPLv3 open source license. Compiled with USE_SSE: fast aligned memory access Compiled with USE_MKL: Intel Math Kernel Library linear algebra Boost version: 1_58 Command line options: /n/groups/price/poru/HSPH_SVN/software/BOLT-LMM_v2.3.1/bolt \ --bed=/n/groups/price/UKBiobank/download_500K/../qc0_500K/chr{1:23}.bed \ --bim=/n/groups/price/UKBiobank/download_500K/../qc0_500K/chr{1:23}.bim \ --fam=/n/groups/price/UKBiobank/download_500K/ukb1404_cal_chr1_v2_CURRENT.fixCol6.fam \ --allowX \ --remove=/n/groups/price/UKBiobank/download_500K/bolt.in_plink_but_not_imputed.FID_IID.976.txt \ --remove=/n/groups/price/UKBiobank/download_500K/../sampleQC/remove.nonWhite.FID_IID.txt \ --exclude=/n/groups/price/UKBiobank/snpQC/allchr.no_phasing.keep_rare.hwe200.txt \ --exclude=/n/groups/price/UKBiobank/snpQC/allchr_maf_lt_1e-5.txt \ --exclude=/n/groups/price/UKBiobank/snpQC/allchr_missing_gt_0.09.txt \ --phenoFile=/n/groups/price/steven/RareVariants/Final/UKB_new_sumstats/UKB_v2.092517.tab \ --phenoCol=blood_HIGH_LIGHT_SCATTER_RETICULOCYTE_COUNT \ --covarFile=/n/groups/price/UKBiobank/app10438assoc/ukb4777.processed_and_post.plinkPCs.tab.gz \ --covarCol=cov_ASSESS_CENTER \ --covarCol=cov_GENO_ARRAY \ --covarCol=cov_SEX \ --covarMaxLevels=30 \ --qCovarCol=cov_AGE \ --qCovarCol=cov_AGE_SQ \ --qCovarCol=PC{1:20} \ --LDscoresFile=/n/groups/price/poru/HSPH_SVN/software/BOLT-LMM_v2.3/tables/LDSCORE.1000G_EUR.tab.gz \ --geneticMapFile=/n/groups/price/poru/HSPH_SVN/software/BOLT-LMM_v2.3/tables/genetic_map_hg19_withX.txt.gz \ --lmmForceNonInf \ --numThreads=8 \ --predBetasFile=bolt_460K_selfRepWhite.blood_HIGH_LIGHT_SCATTER_RETICULOCYTE_COUNT.predbetas.txt.gz \ --statsFile=bolt_460K_selfRepWhite.blood_HIGH_LIGHT_SCATTER_RETICULOCYTE_COUNT.stats.gz \ --verboseStats Setting number of threads to 8 fam: /n/groups/price/UKBiobank/download_500K/ukb1404_cal_chr1_v2_CURRENT.fixCol6.fam bim(s): /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr1.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr2.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr3.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr4.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr5.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr6.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr7.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr8.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr9.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr10.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr11.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr12.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr13.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr14.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr15.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr16.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr17.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr18.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr19.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr20.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr21.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr22.bim /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr23.bim bed(s): /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr1.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr2.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr3.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr4.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr5.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr6.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr7.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr8.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr9.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr10.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr11.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr12.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr13.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr14.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr15.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr16.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr17.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr18.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr19.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr20.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr21.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr22.bed /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr23.bed === Reading genotype data === Total indivs in PLINK data: Nbed = 488377 Reading remove file (indivs to remove): /n/groups/price/UKBiobank/download_500K/bolt.in_plink_but_not_imputed.FID_IID.976.txt Removed 976 individual(s) Reading remove file (indivs to remove): /n/groups/price/UKBiobank/download_500K/../sampleQC/remove.nonWhite.FID_IID.txt WARNING: Unable to find individual to remove: FID IID WARNING: Unable to find individual to remove: 6023494 6023494 WARNING: Unable to find individual to remove: 6022857 6022857 WARNING: Unable to find individual to remove: 6020026 6020026 WARNING: Unable to find individual to remove: 6017119 6017119 Removed 28074 individual(s) WARNING: 1684 individual(s) not found in data set Total indivs stored in memory: N = 459327 Reading bim file #1: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr1.bim Read 63487 snps Reading bim file #2: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr2.bim Read 61966 snps Reading bim file #3: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr3.bim Read 52300 snps Reading bim file #4: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr4.bim Read 47443 snps Reading bim file #5: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr5.bim Read 46314 snps Reading bim file #6: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr6.bim Read 53695 snps Reading bim file #7: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr7.bim Read 42722 snps Reading bim file #8: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr8.bim Read 38591 snps Reading bim file #9: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr9.bim Read 34310 snps Reading bim file #10: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr10.bim Read 38308 snps Reading bim file #11: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr11.bim Read 40824 snps Reading bim file #12: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr12.bim Read 37302 snps Reading bim file #13: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr13.bim Read 26806 snps Reading bim file #14: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr14.bim Read 25509 snps Reading bim file #15: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr15.bim Read 24467 snps Reading bim file #16: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr16.bim Read 28960 snps Reading bim file #17: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr17.bim Read 28835 snps Reading bim file #18: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr18.bim Read 21962 snps Reading bim file #19: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr19.bim Read 26186 snps Reading bim file #20: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr20.bim Read 19959 snps Reading bim file #21: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr21.bim Read 11342 snps Reading bim file #22: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr22.bim Read 12968 snps Reading bim file #23: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr23.bim Read 20214 snps Total snps in PLINK data: Mbed = 804470 Reading exclude file (SNPs to exclude): /n/groups/price/UKBiobank/snpQC/allchr.no_phasing.keep_rare.hwe200.txt Excluded 73451 SNP(s) Reading exclude file (SNPs to exclude): /n/groups/price/UKBiobank/snpQC/allchr_maf_lt_1e-5.txt WARNING: Unable to find SNP to exclude: rs1800865 WARNING: Unable to find SNP to exclude: Affx-89009265 WARNING: Unable to find SNP to exclude: Affx-89005343 WARNING: Unable to find SNP to exclude: Affx-89020964 WARNING: Unable to find SNP to exclude: Affx-89013736 Excluded 16710 SNP(s) WARNING: 342 SNP(s) not found in data set Reading exclude file (SNPs to exclude): /n/groups/price/UKBiobank/snpQC/allchr_missing_gt_0.09.txt WARNING: Unable to find SNP to exclude: rs1800865 WARNING: Unable to find SNP to exclude: Affx-89022776 WARNING: Unable to find SNP to exclude: Affx-89017694 WARNING: Unable to find SNP to exclude: Affx-89018603 WARNING: Unable to find SNP to exclude: Affx-79443721 Excluded 8428 SNP(s) WARNING: 112 SNP(s) not found in data set Breakdown of SNP pre-filtering results: 705881 SNPs to include in model (i.e., GRM) 0 additional non-GRM SNPs loaded 98589 excluded SNPs Filling in genetic map coordinates using reference file: /n/groups/price/poru/HSPH_SVN/software/BOLT-LMM_v2.3/tables/genetic_map_hg19_withX.txt.gz Allocating 705881 x 459328/4 bytes to store genotypes Reading genotypes and performing QC filtering on snps and indivs... Reading bed file #1: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr1.bed Expecting 7751445265 (+3) bytes for 488377 indivs, 63487 snps Reading bed file #2: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr2.bed Expecting 7565738770 (+3) bytes for 488377 indivs, 61966 snps Reading bed file #3: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr3.bed Expecting 6385568500 (+3) bytes for 488377 indivs, 52300 snps Reading bed file #4: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr4.bed Expecting 5792553085 (+3) bytes for 488377 indivs, 47443 snps Reading bed file #5: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr5.bed Expecting 5654707830 (+3) bytes for 488377 indivs, 46314 snps Reading bed file #6: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr6.bed Expecting 6555891025 (+3) bytes for 488377 indivs, 53695 snps Reading bed file #7: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr7.bed Expecting 5216142590 (+3) bytes for 488377 indivs, 42722 snps Reading bed file #8: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr8.bed Expecting 4711768145 (+3) bytes for 488377 indivs, 38591 snps Reading bed file #9: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr9.bed Expecting 4189079450 (+3) bytes for 488377 indivs, 34310 snps Reading bed file #10: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr10.bed Expecting 4677215260 (+3) bytes for 488377 indivs, 38308 snps Reading bed file #11: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr11.bed Expecting 4984406280 (+3) bytes for 488377 indivs, 40824 snps Reading bed file #12: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr12.bed Expecting 4554387690 (+3) bytes for 488377 indivs, 37302 snps Reading bed file #13: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr13.bed Expecting 3272878570 (+3) bytes for 488377 indivs, 26806 snps Reading bed file #14: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr14.bed Expecting 3114521355 (+3) bytes for 488377 indivs, 25509 snps Reading bed file #15: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr15.bed Expecting 2987298365 (+3) bytes for 488377 indivs, 24467 snps Reading bed file #16: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr16.bed Expecting 3535871200 (+3) bytes for 488377 indivs, 28960 snps Reading bed file #17: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr17.bed Expecting 3520609325 (+3) bytes for 488377 indivs, 28835 snps Reading bed file #18: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr18.bed Expecting 2681450390 (+3) bytes for 488377 indivs, 21962 snps Reading bed file #19: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr19.bed Expecting 3197179670 (+3) bytes for 488377 indivs, 26186 snps Reading bed file #20: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr20.bed Expecting 2436894105 (+3) bytes for 488377 indivs, 19959 snps Reading bed file #21: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr21.bed Expecting 1384801490 (+3) bytes for 488377 indivs, 11342 snps Reading bed file #22: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr22.bed Expecting 1583327960 (+3) bytes for 488377 indivs, 12968 snps Reading bed file #23: /n/groups/price/UKBiobank/download_500K/../qc0_500K/chr23.bed Expecting 2468028330 (+3) bytes for 488377 indivs, 20214 snps Total indivs after QC: 459327 Total post-QC SNPs: M = 705881 Variance component 1: 705881 post-QC SNPs (name: 'modelSnps') Time for SnpData setup = 3024.82 sec === Reading phenotype and covariate data === WARNING: Ignoring indiv not in genotype data: FID=6018780, IID=6018780 WARNING: Ignoring indiv not in genotype data: FID=6012488, IID=6012488 WARNING: Ignoring indiv not in genotype data: FID=5998913, IID=5998913 WARNING: Ignoring indiv not in genotype data: FID=5989416, IID=5989416 WARNING: Ignoring indiv not in genotype data: FID=5985954, IID=5985954 Read data for 460238 indivs (ignored 914 without genotypes) from: /n/groups/price/UKBiobank/app10438assoc/ukb4777.processed_and_post.plinkPCs.tab.gz WARNING: Ignoring indiv not in genotype data: FID=1000129, IID=1000129 WARNING: Ignoring indiv not in genotype data: FID=1000170, IID=1000170 WARNING: Ignoring indiv not in genotype data: FID=1000224, IID=1000224 WARNING: Ignoring indiv not in genotype data: FID=1000362, IID=1000362 WARNING: Ignoring indiv not in genotype data: FID=1000379, IID=1000379 Read data for 502655 indivs (ignored 43331 without genotypes) from: /n/groups/price/steven/RareVariants/Final/UKB_new_sumstats/UKB_v2.092517.tab Number of indivs with no missing phenotype(s) to use: 437490 NOTE: Using all-1s vector (constant term) in addition to specified covariates Using categorical covariate: cov_ASSESS_CENTER (adding level 10003) Using categorical covariate: cov_ASSESS_CENTER (adding level 11001) Using categorical covariate: cov_ASSESS_CENTER (adding level 11002) Using categorical covariate: cov_ASSESS_CENTER (adding level 11003) Using categorical covariate: cov_ASSESS_CENTER (adding level 11004) Using categorical covariate: cov_ASSESS_CENTER (adding level 11005) Using categorical covariate: cov_ASSESS_CENTER (adding level 11006) Using categorical covariate: cov_ASSESS_CENTER (adding level 11007) Using categorical covariate: cov_ASSESS_CENTER (adding level 11008) Using categorical covariate: cov_ASSESS_CENTER (adding level 11009) Using categorical covariate: cov_ASSESS_CENTER (adding level 11010) Using categorical covariate: cov_ASSESS_CENTER (adding level 11011) Using categorical covariate: cov_ASSESS_CENTER (adding level 11012) Using categorical covariate: cov_ASSESS_CENTER (adding level 11013) Using categorical covariate: cov_ASSESS_CENTER (adding level 11014) Using categorical covariate: cov_ASSESS_CENTER (adding level 11016) Using categorical covariate: cov_ASSESS_CENTER (adding level 11017) Using categorical covariate: cov_ASSESS_CENTER (adding level 11018) Using categorical covariate: cov_ASSESS_CENTER (adding level 11020) Using categorical covariate: cov_ASSESS_CENTER (adding level 11021) Using categorical covariate: cov_ASSESS_CENTER (adding level 11022) Using categorical covariate: cov_ASSESS_CENTER (adding level 11023) WARNING: Covariate cov_ASSESS_CENTER has a large number of distinct values (23) Should this covariate be quantitative rather than categorical? Using categorical covariate: cov_GENO_ARRAY (adding level 1) Using categorical covariate: cov_GENO_ARRAY (adding level 2) Using categorical covariate: cov_SEX (adding level 0) Using categorical covariate: cov_SEX (adding level 1) Using quantitative covariate: cov_AGE Using quantitative covariate: cov_AGE_SQ Using quantitative covariate: PC1 Using quantitative covariate: PC2 Using quantitative covariate: PC3 Using quantitative covariate: PC4 Using quantitative covariate: PC5 Using quantitative covariate: PC6 Using quantitative covariate: PC7 Using quantitative covariate: PC8 Using quantitative covariate: PC9 Using quantitative covariate: PC10 Using quantitative covariate: PC11 Using quantitative covariate: PC12 Using quantitative covariate: PC13 Using quantitative covariate: PC14 Using quantitative covariate: PC15 Using quantitative covariate: PC16 Using quantitative covariate: PC17 Using quantitative covariate: PC18 Using quantitative covariate: PC19 Using quantitative covariate: PC20 Using quantitative covariate: CONST_ALL_ONES Number of individuals used in analysis: Nused = 437490 Singular values of covariate matrix: S[0] = 2.25232e+06 S[1] = 4726.15 S[2] = 465.987 S[3] = 291.401 S[4] = 198.976 S[5] = 193.486 S[6] = 182.953 S[7] = 174.33 S[8] = 169.444 S[9] = 163.025 S[10] = 161.339 S[11] = 151.533 S[12] = 142.47 S[13] = 139.807 S[14] = 138.828 S[15] = 135.098 S[16] = 132.421 S[17] = 130.183 S[18] = 126.148 S[19] = 116.071 S[20] = 112.038 S[21] = 99.7138 S[22] = 43.1682 S[23] = 23.9571 S[24] = 19.1553 S[25] = 0.976032 S[26] = 0.975844 S[27] = 0.975817 S[28] = 0.975638 S[29] = 0.97552 S[30] = 0.975418 S[31] = 0.975288 S[32] = 0.975185 S[33] = 0.975033 S[34] = 0.97492 S[35] = 0.974841 S[36] = 0.974668 S[37] = 0.97452 S[38] = 0.974375 S[39] = 0.974206 S[40] = 0.973957 S[41] = 0.973857 S[42] = 0.973826 S[43] = 0.95551 S[44] = 0.880951 S[45] = 5.21035e-12 S[46] = 5.36877e-13 S[47] = 3.87544e-13 S[48] = 0 Total covariate vectors: C = 49 Total independent covariate vectors: Cindep = 45 === Initializing Bolt object: projecting and normalizing SNPs === Number of chroms with >= 1 good SNP: 23 Average norm of projected SNPs: 434940.843218 Dimension of all-1s proj space (Nused-1): 437489 Time for covariate data setup + Bolt initialization = 4465.02 sec Phenotype 1: N = 437490 mean = -0.0135397 std = 0.998688 === Computing linear regression (LINREG) stats === Time for computing LINREG stats = 541.107 sec === Estimating variance parameters === Using CGtol of 0.005 for this step Using default number of random trials: 3 (for Nused = 437490) Estimating MC scaling f_REML at log(delta) = 1.09287, h2 = 0.25... Batch-solving 4 systems of equations using conjugate gradient iteration iter 1: time=403.66 rNorms/orig: (0.6,0.7) res2s: 713959..145608 iter 2: time=392.23 rNorms/orig: (0.5,0.6) res2s: 845223..188601 iter 3: time=377.93 rNorms/orig: (0.3,0.4) res2s: 975577..222145 iter 4: time=383.12 rNorms/orig: (0.2,0.2) res2s: 1.02069e+06..237776 iter 5: time=382.84 rNorms/orig: (0.1,0.2) res2s: 1.04682e+06..244726 iter 6: time=398.76 rNorms/orig: (0.09,0.1) res2s: 1.05882e+06..248731 iter 7: time=398.33 rNorms/orig: (0.05,0.06) res2s: 1.06488e+06..250613 iter 8: time=405.22 rNorms/orig: (0.04,0.04) res2s: 1.06785e+06..251435 iter 9: time=408.91 rNorms/orig: (0.02,0.03) res2s: 1.06918e+06..251791 iter 10: time=425.29 rNorms/orig: (0.01,0.02) res2s: 1.06981e+06..251978 iter 11: time=450.83 rNorms/orig: (0.01,0.01) res2s: 1.0701e+06..252065 iter 12: time=438.00 rNorms/orig: (0.006,0.007) res2s: 1.07023e+06..252103 iter 13: time=412.27 rNorms/orig: (0.004,0.005) res2s: 1.07028e+06..252117 Converged at iter 13: rNorms/orig all < CGtol=0.005 Time breakdown: dgemm = 48.0%, memory/overhead = 52.0% MCscaling: logDelta = 1.09, h2 = 0.250, f = 0.0819186 Estimating MC scaling f_REML at log(delta) = -0.00574095, h2 = 0.5... Batch-solving 4 systems of equations using conjugate gradient iteration iter 1: time=393.43 rNorms/orig: (1,1) res2s: 76297.7..42893.1 iter 2: time=387.88 rNorms/orig: (1,1) res2s: 109959..68466.6 iter 3: time=386.71 rNorms/orig: (0.8,0.9) res2s: 165836..100015 iter 4: time=391.27 rNorms/orig: (0.6,0.7) res2s: 199549..123041 iter 5: time=402.08 rNorms/orig: (0.5,0.5) res2s: 228087..137832 iter 6: time=431.51 rNorms/orig: (0.4,0.4) res2s: 247571..149980 iter 7: time=448.58 rNorms/orig: (0.3,0.3) res2s: 261851..158341 iter 8: time=421.48 rNorms/orig: (0.2,0.2) res2s: 272029..163626 iter 9: time=423.28 rNorms/orig: (0.2,0.2) res2s: 278587..166876 iter 10: time=431.04 rNorms/orig: (0.1,0.1) res2s: 283076..169299 iter 11: time=420.37 rNorms/orig: (0.1,0.1) res2s: 286142..170914 iter 12: time=407.99 rNorms/orig: (0.08,0.09) res2s: 288065..171970 iter 13: time=401.46 rNorms/orig: (0.07,0.07) res2s: 289221..172528 iter 14: time=390.66 rNorms/orig: (0.05,0.05) res2s: 290047..172947 iter 15: time=386.15 rNorms/orig: (0.04,0.04) res2s: 290574..173209 iter 16: time=402.00 rNorms/orig: (0.03,0.03) res2s: 290891..173369 iter 17: time=403.47 rNorms/orig: (0.02,0.03) res2s: 291078..173472 iter 18: time=404.86 rNorms/orig: (0.02,0.02) res2s: 291222..173538 iter 19: time=403.04 rNorms/orig: (0.01,0.01) res2s: 291306..173579 iter 20: time=392.12 rNorms/orig: (0.01,0.01) res2s: 291357..173606 iter 21: time=405.93 rNorms/orig: (0.008,0.009) res2s: 291388..173621 iter 22: time=392.00 rNorms/orig: (0.006,0.007) res2s: 291406..173629 iter 23: time=406.39 rNorms/orig: (0.005,0.005) res2s: 291418..173635 iter 24: time=405.94 rNorms/orig: (0.004,0.004) res2s: 291425..173639 Converged at iter 24: rNorms/orig all < CGtol=0.005 Time breakdown: dgemm = 48.3%, memory/overhead = 51.7% MCscaling: logDelta = -0.01, h2 = 0.500, f = -0.239839 Estimating MC scaling f_REML at log(delta) = 0.813168, h2 = 0.305995... Batch-solving 4 systems of equations using conjugate gradient iteration iter 1: time=402.04 rNorms/orig: (0.7,0.8) res2s: 433527..112841 iter 2: time=400.42 rNorms/orig: (0.7,0.8) res2s: 539607..154035 iter 3: time=408.16 rNorms/orig: (0.4,0.5) res2s: 661029..190530 iter 4: time=402.59 rNorms/orig: (0.3,0.3) res2s: 709677..209788 iter 5: time=389.48 rNorms/orig: (0.2,0.2) res2s: 740816..219294 iter 6: time=408.41 rNorms/orig: (0.1,0.2) res2s: 756794..225371 iter 7: time=416.87 rNorms/orig: (0.09,0.1) res2s: 765770..228572 iter 8: time=408.39 rNorms/orig: (0.06,0.07) res2s: 770690..230135 iter 9: time=401.95 rNorms/orig: (0.05,0.05) res2s: 773132..230890 iter 10: time=395.34 rNorms/orig: (0.03,0.03) res2s: 774436..231331 iter 11: time=406.07 rNorms/orig: (0.02,0.02) res2s: 775114..231559 iter 12: time=400.85 rNorms/orig: (0.01,0.02) res2s: 775445..231673 iter 13: time=381.70 rNorms/orig: (0.01,0.01) res2s: 775598..231719 iter 14: time=401.22 rNorms/orig: (0.006,0.007) res2s: 775685..231747 iter 15: time=406.66 rNorms/orig: (0.004,0.005) res2s: 775727..231760 Converged at iter 15: rNorms/orig all < CGtol=0.005 Time breakdown: dgemm = 48.0%, memory/overhead = 52.0% MCscaling: logDelta = 0.81, h2 = 0.306, f = 0.00144361 Secant iteration for h2 estimation converged in 1 steps Estimated (pseudo-)heritability: h2g = 0.306 To more precisely estimate variance parameters and s.e., use --reml Variance params: sigma^2_K = 0.299519, logDelta = 0.813168, f = 0.00144361 Time for fitting variance components = 21806.6 sec === Computing mixed model assoc stats (inf. model) === Selected 30 SNPs for computation of prospective stat Tried 30; threw out 0 with GRAMMAR chisq > 5 Assigning SNPs to 23 chunks for leave-out analysis Each chunk is excluded when testing SNPs belonging to the chunk Batch-solving 53 systems of equations using conjugate gradient iteration iter 1: time=768.21 rNorms/orig: (0.5,0.9) res2s: 113070..205469 iter 2: time=783.56 rNorms/orig: (0.5,0.8) res2s: 155109..244864 iter 3: time=793.76 rNorms/orig: (0.3,0.5) res2s: 192859..264555 iter 4: time=734.06 rNorms/orig: (0.2,0.4) res2s: 213847..273494 iter 5: time=729.10 rNorms/orig: (0.2,0.2) res2s: 224203..279308 iter 6: time=729.62 rNorms/orig: (0.1,0.2) res2s: 230883..282112 iter 7: time=734.20 rNorms/orig: (0.07,0.1) res2s: 234723..283672 iter 8: time=734.76 rNorms/orig: (0.05,0.08) res2s: 236543..284561 iter 9: time=736.39 rNorms/orig: (0.03,0.06) res2s: 237422..284994 iter 10: time=742.77 rNorms/orig: (0.02,0.04) res2s: 238010..285228 iter 11: time=737.49 rNorms/orig: (0.02,0.03) res2s: 238293..285349 iter 12: time=743.51 rNorms/orig: (0.01,0.02) res2s: 238432..285411 iter 13: time=740.17 rNorms/orig: (0.006,0.01) res2s: 238500..285441 iter 14: time=734.52 rNorms/orig: (0.004,0.009) res2s: 238539..285458 iter 15: time=769.26 rNorms/orig: (0.003,0.006) res2s: 238558..285465 iter 16: time=778.17 rNorms/orig: (0.002,0.005) res2s: 238567..285469 iter 17: time=759.75 rNorms/orig: (0.001,0.003) res2s: 238572..285471 iter 18: time=732.29 rNorms/orig: (0.0007,0.002) res2s: 238574..285472 iter 19: time=760.45 rNorms/orig: (0.0004,0.001) res2s: 238576..285472 iter 20: time=749.75 rNorms/orig: (0.0003,0.0009) res2s: 238576..285472 iter 21: time=746.13 rNorms/orig: (0.0002,0.0006) res2s: 238577..285472 iter 22: time=749.66 rNorms/orig: (0.0001,0.0004) res2s: 238577..285473 Converged at iter 22: rNorms/orig all < CGtol=0.0005 Time breakdown: dgemm = 72.6%, memory/overhead = 27.4% AvgPro: 2.014 AvgRetro: 1.984 Calibration: 1.015 (0.002) (30 SNPs) Ratio of medians: 1.013 Median of ratios: 1.013 Time for computing infinitesimal model assoc stats = 16966.9 sec === Estimating chip LD Scores using 400 indivs === Time for estimating chip LD Scores = 18.2173 sec === Reading LD Scores for calibration of Bayesian assoc stats === Looking up LD Scores... Looking for column header 'SNP': column number = 1 Looking for column header 'LDSCORE': column number = 5 Found LD Scores for 601289/705881 SNPs Estimating inflation of LINREG chisq stats using MLMe as reference... Filtering to SNPs with chisq stats, LD Scores, and MAF > 0.01 # of SNPs passing filters before outlier removal: 579570/705881 Masking windows around outlier snps (chisq > 437.5) # of SNPs remaining after outlier window removal: 572077/579570 Intercept of LD Score regression for ref stats: 1.157 (0.017) Estimated attenuation: 0.145 (0.016) Intercept of LD Score regression for cur stats: 1.157 (0.016) Calibration factor (ref/cur) to multiply by: 1.000 (0.003) LINREG intercept inflation = 1.00012 NOTE: LINREG stats in output are NOT corrected for estimated inflation === Estimating mixture parameters by cross-validation === Setting maximum number of iterations to 250 for this step Max CV folds to compute = 1 (to have > 10000 samples) ====> Starting CV fold 1 <==== NOTE: Using all-1s vector (constant term) in addition to specified covariates Using categorical covariate: cov_ASSESS_CENTER (adding level 10003) Using categorical covariate: cov_ASSESS_CENTER (adding level 11001) Using categorical covariate: cov_ASSESS_CENTER (adding level 11002) Using categorical covariate: cov_ASSESS_CENTER (adding level 11003) Using categorical covariate: cov_ASSESS_CENTER (adding level 11004) Using categorical covariate: cov_ASSESS_CENTER (adding level 11005) Using categorical covariate: cov_ASSESS_CENTER (adding level 11006) Using categorical covariate: cov_ASSESS_CENTER (adding level 11007) Using categorical covariate: cov_ASSESS_CENTER (adding level 11008) Using categorical covariate: cov_ASSESS_CENTER (adding level 11009) Using categorical covariate: cov_ASSESS_CENTER (adding level 11010) Using categorical covariate: cov_ASSESS_CENTER (adding level 11011) Using categorical covariate: cov_ASSESS_CENTER (adding level 11012) Using categorical covariate: cov_ASSESS_CENTER (adding level 11013) Using categorical covariate: cov_ASSESS_CENTER (adding level 11014) Using categorical covariate: cov_ASSESS_CENTER (adding level 11016) Using categorical covariate: cov_ASSESS_CENTER (adding level 11017) Using categorical covariate: cov_ASSESS_CENTER (adding level 11018) Using categorical covariate: cov_ASSESS_CENTER (adding level 11020) Using categorical covariate: cov_ASSESS_CENTER (adding level 11021) Using categorical covariate: cov_ASSESS_CENTER (adding level 11022) Using categorical covariate: cov_ASSESS_CENTER (adding level 11023) WARNING: Covariate cov_ASSESS_CENTER has a large number of distinct values (23) Should this covariate be quantitative rather than categorical? Using categorical covariate: cov_GENO_ARRAY (adding level 1) Using categorical covariate: cov_GENO_ARRAY (adding level 2) Using categorical covariate: cov_SEX (adding level 0) Using categorical covariate: cov_SEX (adding level 1) Using quantitative covariate: cov_AGE Using quantitative covariate: cov_AGE_SQ Using quantitative covariate: PC1 Using quantitative covariate: PC2 Using quantitative covariate: PC3 Using quantitative covariate: PC4 Using quantitative covariate: PC5 Using quantitative covariate: PC6 Using quantitative covariate: PC7 Using quantitative covariate: PC8 Using quantitative covariate: PC9 Using quantitative covariate: PC10 Using quantitative covariate: PC11 Using quantitative covariate: PC12 Using quantitative covariate: PC13 Using quantitative covariate: PC14 Using quantitative covariate: PC15 Using quantitative covariate: PC16 Using quantitative covariate: PC17 Using quantitative covariate: PC18 Using quantitative covariate: PC19 Using quantitative covariate: PC20 Using quantitative covariate: CONST_ALL_ONES Number of individuals used in analysis: Nused = 349992 Singular values of covariate matrix: S[0] = 2.01513e+06 S[1] = 4226.21 S[2] = 416.804 S[3] = 260.641 S[4] = 177.981 S[5] = 173.043 S[6] = 163.577 S[7] = 155.84 S[8] = 151.54 S[9] = 145.913 S[10] = 144.296 S[11] = 135.561 S[12] = 127.52 S[13] = 125.038 S[14] = 124.072 S[15] = 120.807 S[16] = 118.474 S[17] = 116.534 S[18] = 112.922 S[19] = 103.686 S[20] = 100.133 S[21] = 89.2205 S[22] = 38.7777 S[23] = 21.5111 S[24] = 17.1316 S[25] = 0.874746 S[26] = 0.874323 S[27] = 0.87379 S[28] = 0.87366 S[29] = 0.873369 S[30] = 0.872986 S[31] = 0.872733 S[32] = 0.872286 S[33] = 0.872129 S[34] = 0.87177 S[35] = 0.871555 S[36] = 0.871271 S[37] = 0.871052 S[38] = 0.870645 S[39] = 0.870308 S[40] = 0.869593 S[41] = 0.869069 S[42] = 0.868664 S[43] = 0.858175 S[44] = 0.788233 S[45] = 5.71244e-12 S[46] = 1.22551e-12 S[47] = 1.48235e-13 S[48] = 0 Total covariate vectors: C = 49 Total independent covariate vectors: Cindep = 45 === Initializing Bolt object: projecting and normalizing SNPs === Number of chroms with >= 1 good SNP: 23 Average norm of projected SNPs: 347942.453699 Dimension of all-1s proj space (Nused-1): 349991 Beginning variational Bayes iter 1: time=747.56 for 18 active reps iter 2: time=483.81 for 18 active reps approxLL diffs: (11415.43,14097.42) iter 3: time=488.86 for 18 active reps approxLL diffs: (1141.85,2580.68) iter 4: time=504.44 for 18 active reps approxLL diffs: (225.07,1007.82) iter 5: time=492.71 for 18 active reps approxLL diffs: (68.81,568.82) iter 6: time=490.26 for 18 active reps approxLL diffs: (28.25,332.83) iter 7: time=495.46 for 18 active reps approxLL diffs: (14.15,214.53) iter 8: time=482.51 for 18 active reps approxLL diffs: (8.07,125.12) iter 9: time=478.26 for 18 active reps approxLL diffs: (4.97,98.55) iter 10: time=467.98 for 18 active reps approxLL diffs: (3.28,64.82) iter 11: time=473.79 for 18 active reps approxLL diffs: (2.30,48.22) iter 12: time=471.99 for 18 active reps approxLL diffs: (1.69,39.28) iter 13: time=474.81 for 18 active reps approxLL diffs: (1.28,32.38) iter 14: time=473.83 for 18 active reps approxLL diffs: (0.98,23.18) iter 15: time=469.20 for 18 active reps approxLL diffs: (0.77,18.86) iter 16: time=464.79 for 18 active reps approxLL diffs: (0.63,17.52) iter 17: time=466.41 for 18 active reps approxLL diffs: (0.53,15.45) iter 18: time=466.95 for 18 active reps approxLL diffs: (0.45,14.52) iter 19: time=465.42 for 18 active reps approxLL diffs: (0.37,12.17) iter 20: time=467.31 for 18 active reps approxLL diffs: (0.30,8.12) iter 21: time=464.50 for 18 active reps approxLL diffs: (0.25,6.03) iter 22: time=462.31 for 18 active reps approxLL diffs: (0.21,5.38) iter 23: time=462.19 for 18 active reps approxLL diffs: (0.17,8.32) iter 24: time=462.16 for 18 active reps approxLL diffs: (0.14,5.65) iter 25: time=461.16 for 18 active reps approxLL diffs: (0.11,3.36) iter 26: time=460.49 for 18 active reps approxLL diffs: (0.10,4.18) iter 27: time=459.31 for 18 active reps approxLL diffs: (0.09,6.40) iter 28: time=460.64 for 18 active reps approxLL diffs: (0.08,4.45) iter 29: time=459.91 for 18 active reps approxLL diffs: (0.06,5.15) iter 30: time=461.68 for 18 active reps approxLL diffs: (0.05,7.95) iter 31: time=470.54 for 18 active reps approxLL diffs: (0.04,8.20) iter 32: time=468.97 for 18 active reps approxLL diffs: (0.04,3.77) iter 33: time=478.74 for 18 active reps approxLL diffs: (0.03,2.98) iter 34: time=490.28 for 18 active reps approxLL diffs: (0.03,2.60) iter 35: time=478.08 for 18 active reps approxLL diffs: (0.02,2.06) iter 36: time=459.84 for 18 active reps approxLL diffs: (0.02,1.72) iter 37: time=467.06 for 18 active reps approxLL diffs: (0.02,1.65) iter 38: time=479.20 for 18 active reps approxLL diffs: (0.02,1.85) iter 39: time=488.19 for 18 active reps approxLL diffs: (0.01,2.42) iter 40: time=479.91 for 18 active reps approxLL diffs: (0.01,2.74) iter 41: time=481.19 for 18 active reps approxLL diffs: (0.01,4.80) iter 42: time=482.81 for 18 active reps approxLL diffs: (0.01,3.20) iter 43: time=432.08 for 16 active reps approxLL diffs: (0.01,2.20) iter 44: time=429.44 for 14 active reps approxLL diffs: (0.01,0.45) iter 45: time=419.15 for 13 active reps approxLL diffs: (0.01,0.28) iter 46: time=398.41 for 12 active reps approxLL diffs: (0.01,0.23) iter 47: time=408.06 for 11 active reps approxLL diffs: (0.01,0.78) iter 48: time=401.40 for 10 active reps approxLL diffs: (0.01,2.51) iter 49: time=390.23 for 10 active reps approxLL diffs: (0.01,0.84) iter 50: time=384.47 for 10 active reps approxLL diffs: (0.01,0.20) iter 51: time=389.22 for 10 active reps approxLL diffs: (0.01,0.13) iter 52: time=380.34 for 10 active reps approxLL diffs: (0.01,0.12) iter 53: time=324.59 for 8 active reps approxLL diffs: (0.01,0.16) iter 54: time=318.05 for 8 active reps approxLL diffs: (0.01,0.53) iter 55: time=323.95 for 8 active reps approxLL diffs: (0.01,1.21) iter 56: time=327.55 for 8 active reps approxLL diffs: (0.01,0.62) iter 57: time=354.60 for 7 active reps approxLL diffs: (0.02,0.23) iter 58: time=342.96 for 7 active reps approxLL diffs: (0.01,0.32) iter 59: time=326.94 for 6 active reps approxLL diffs: (0.01,0.27) iter 60: time=314.03 for 5 active reps approxLL diffs: (0.01,0.10) iter 61: time=297.83 for 5 active reps approxLL diffs: (0.01,0.10) iter 62: time=290.21 for 3 active reps approxLL diffs: (0.01,0.22) iter 63: time=271.83 for 2 active reps approxLL diffs: (0.01,0.29) iter 64: time=274.18 for 2 active reps approxLL diffs: (0.02,0.11) iter 65: time=282.28 for 2 active reps approxLL diffs: (0.04,0.05) iter 66: time=272.78 for 2 active reps approxLL diffs: (0.08,0.13) iter 67: time=273.03 for 2 active reps approxLL diffs: (0.13,0.43) iter 68: time=272.03 for 2 active reps approxLL diffs: (0.14,0.86) iter 69: time=271.18 for 2 active reps approxLL diffs: (0.09,1.09) iter 70: time=269.89 for 2 active reps approxLL diffs: (0.09,0.86) iter 71: time=272.87 for 2 active reps approxLL diffs: (0.11,0.14) iter 72: time=274.99 for 2 active reps approxLL diffs: (0.09,0.11) iter 73: time=277.98 for 2 active reps approxLL diffs: (0.04,0.38) iter 74: time=274.40 for 2 active reps approxLL diffs: (0.02,0.69) iter 75: time=276.36 for 2 active reps approxLL diffs: (0.02,0.40) iter 76: time=284.92 for 2 active reps approxLL diffs: (0.04,0.12) iter 77: time=279.75 for 2 active reps approxLL diffs: (0.03,0.09) iter 78: time=280.00 for 2 active reps approxLL diffs: (0.01,0.14) iter 79: time=231.57 for 1 active reps approxLL diffs: (0.13,0.13) iter 80: time=232.31 for 1 active reps approxLL diffs: (0.07,0.07) iter 81: time=233.61 for 1 active reps approxLL diffs: (0.02,0.02) iter 82: time=221.56 for 1 active reps approxLL diffs: (0.01,0.01) Converged at iter 82: approxLL diffs each have been < LLtol=0.01 Time breakdown: dgemm = 75.7%, memory/overhead = 24.3% Computing predictions on left-out cross-validation fold Time for computing predictions = 9289.58 sec Average PVEs obtained by param pairs tested (high to low): f2=0.5, p=0.01: 0.171890 f2=0.5, p=0.02: 0.171403 f2=0.3, p=0.02: 0.170932 ... f2=0.5, p=0.5: 0.121329 Detailed CV fold results: Absolute prediction MSE baseline (covariates only): 0.995342 Absolute prediction MSE using standard LMM: 0.874578 Absolute prediction MSE, fold-best f2=0.5, p=0.01: 0.824252 Absolute pred MSE using f2=0.5, p=0.5: 0.874578 Absolute pred MSE using f2=0.5, p=0.2: 0.857100 Absolute pred MSE using f2=0.5, p=0.1: 0.842163 Absolute pred MSE using f2=0.5, p=0.05: 0.831048 Absolute pred MSE using f2=0.5, p=0.02: 0.824737 Absolute pred MSE using f2=0.5, p=0.01: 0.824252 Absolute pred MSE using f2=0.3, p=0.5: 0.868551 Absolute pred MSE using f2=0.3, p=0.2: 0.846858 Absolute pred MSE using f2=0.3, p=0.1: 0.833223 Absolute pred MSE using f2=0.3, p=0.05: 0.826213 Absolute pred MSE using f2=0.3, p=0.02: 0.825206 Absolute pred MSE using f2=0.3, p=0.01: 0.826436 Absolute pred MSE using f2=0.1, p=0.5: 0.860772 Absolute pred MSE using f2=0.1, p=0.2: 0.839040 Absolute pred MSE using f2=0.1, p=0.1: 0.829104 Absolute pred MSE using f2=0.1, p=0.05: 0.828218 Absolute pred MSE using f2=0.1, p=0.02: 0.833326 Absolute pred MSE using f2=0.1, p=0.01: 0.835763 ====> End CV fold 1: 18 remaining param pair(s) <==== Estimated proportion of variance explained using inf model: 0.121 Relative improvement in prediction MSE using non-inf model: 0.058 Optimal mixture parameters according to CV: f2 = 0.5, p = 0.01 Time for estimating mixture parameters = 46587.6 sec === Computing Bayesian mixed model assoc stats with mixture prior === Assigning SNPs to 23 chunks for leave-out analysis Each chunk is excluded when testing SNPs belonging to the chunk Beginning variational Bayes iter 1: time=717.98 for 23 active reps iter 2: time=477.96 for 23 active reps approxLL diffs: (15968.90,17297.70) iter 3: time=472.53 for 23 active reps approxLL diffs: (2965.17,3248.53) iter 4: time=473.74 for 23 active reps approxLL diffs: (1082.40,1193.98) iter 5: time=475.84 for 23 active reps approxLL diffs: (526.27,598.72) iter 6: time=476.22 for 23 active reps approxLL diffs: (302.42,368.93) iter 7: time=483.63 for 23 active reps approxLL diffs: (176.86,240.62) iter 8: time=470.65 for 23 active reps approxLL diffs: (115.86,150.77) iter 9: time=471.02 for 23 active reps approxLL diffs: (79.23,101.70) iter 10: time=472.08 for 23 active reps approxLL diffs: (55.28,76.77) iter 11: time=465.28 for 23 active reps approxLL diffs: (40.87,56.41) iter 12: time=458.33 for 23 active reps approxLL diffs: (22.41,46.33) iter 13: time=469.07 for 23 active reps approxLL diffs: (18.32,39.70) iter 14: time=466.76 for 23 active reps approxLL diffs: (16.29,30.99) iter 15: time=467.72 for 23 active reps approxLL diffs: (15.14,30.09) iter 16: time=492.20 for 23 active reps approxLL diffs: (12.01,23.89) iter 17: time=515.55 for 23 active reps approxLL diffs: (9.95,17.35) iter 18: time=528.31 for 23 active reps approxLL diffs: (7.29,18.61) iter 19: time=527.30 for 23 active reps approxLL diffs: (4.93,14.11) iter 20: time=525.69 for 23 active reps approxLL diffs: (3.20,12.21) iter 21: time=550.66 for 23 active reps approxLL diffs: (3.05,8.43) iter 22: time=555.04 for 23 active reps approxLL diffs: (1.78,6.71) iter 23: time=551.89 for 23 active reps approxLL diffs: (1.55,6.40) iter 24: time=548.40 for 23 active reps approxLL diffs: (1.11,6.52) iter 25: time=554.15 for 23 active reps approxLL diffs: (1.26,7.10) iter 26: time=534.01 for 23 active reps approxLL diffs: (1.46,8.18) iter 27: time=537.66 for 23 active reps approxLL diffs: (1.07,12.54) iter 28: time=537.77 for 23 active reps approxLL diffs: (0.87,5.63) iter 29: time=538.37 for 23 active reps approxLL diffs: (0.58,5.50) iter 30: time=537.34 for 23 active reps approxLL diffs: (0.42,7.78) iter 31: time=536.37 for 23 active reps approxLL diffs: (0.28,4.93) iter 32: time=532.16 for 23 active reps approxLL diffs: (0.25,5.77) iter 33: time=532.86 for 23 active reps approxLL diffs: (0.22,5.85) iter 34: time=544.16 for 23 active reps approxLL diffs: (0.11,4.28) iter 35: time=536.40 for 23 active reps approxLL diffs: (0.08,4.67) iter 36: time=538.43 for 23 active reps approxLL diffs: (0.05,3.03) iter 37: time=542.24 for 23 active reps approxLL diffs: (0.04,2.47) iter 38: time=537.59 for 23 active reps approxLL diffs: (0.04,3.32) iter 39: time=523.40 for 23 active reps approxLL diffs: (0.04,4.30) iter 40: time=520.49 for 23 active reps approxLL diffs: (0.04,5.08) iter 41: time=522.60 for 23 active reps approxLL diffs: (0.04,1.84) iter 42: time=525.37 for 23 active reps approxLL diffs: (0.02,3.64) iter 43: time=525.17 for 23 active reps approxLL diffs: (0.02,2.91) iter 44: time=538.92 for 23 active reps approxLL diffs: (0.02,3.65) iter 45: time=548.80 for 23 active reps approxLL diffs: (0.03,3.08) iter 46: time=546.20 for 23 active reps approxLL diffs: (0.01,2.26) iter 47: time=547.98 for 23 active reps approxLL diffs: (0.00,4.13) iter 48: time=512.52 for 22 active reps approxLL diffs: (0.01,3.76) iter 49: time=510.82 for 22 active reps approxLL diffs: (0.01,2.27) iter 50: time=482.18 for 21 active reps approxLL diffs: (0.01,2.27) iter 51: time=460.75 for 20 active reps approxLL diffs: (0.02,2.88) iter 52: time=460.05 for 20 active reps approxLL diffs: (0.02,2.46) iter 53: time=454.37 for 20 active reps approxLL diffs: (0.01,2.84) iter 54: time=473.67 for 19 active reps approxLL diffs: (0.01,1.80) iter 55: time=474.41 for 19 active reps approxLL diffs: (0.01,2.86) iter 56: time=410.16 for 16 active reps approxLL diffs: (0.01,3.38) iter 57: time=409.84 for 16 active reps approxLL diffs: (0.01,1.02) iter 58: time=420.82 for 15 active reps approxLL diffs: (0.02,1.20) iter 59: time=420.38 for 15 active reps approxLL diffs: (0.03,2.17) iter 60: time=423.69 for 15 active reps approxLL diffs: (0.02,3.92) iter 61: time=423.39 for 15 active reps approxLL diffs: (0.00,1.74) iter 62: time=399.34 for 14 active reps approxLL diffs: (0.01,0.90) iter 63: time=382.81 for 13 active reps approxLL diffs: (0.01,0.53) iter 64: time=381.74 for 11 active reps approxLL diffs: (0.01,0.33) iter 65: time=354.90 for 10 active reps approxLL diffs: (0.01,0.29) iter 66: time=341.34 for 9 active reps approxLL diffs: (0.01,0.36) iter 67: time=345.72 for 9 active reps approxLL diffs: (0.01,0.43) iter 68: time=354.36 for 9 active reps approxLL diffs: (0.01,1.08) iter 69: time=352.84 for 9 active reps approxLL diffs: (0.01,1.52) iter 70: time=299.79 for 8 active reps approxLL diffs: (0.01,1.47) iter 71: time=324.45 for 7 active reps approxLL diffs: (0.01,0.16) iter 72: time=328.37 for 7 active reps approxLL diffs: (0.00,0.31) iter 73: time=305.86 for 6 active reps approxLL diffs: (0.01,1.64) iter 74: time=302.98 for 6 active reps approxLL diffs: (0.02,1.27) iter 75: time=302.74 for 6 active reps approxLL diffs: (0.02,0.10) iter 76: time=306.22 for 6 active reps approxLL diffs: (0.00,0.11) iter 77: time=267.48 for 4 active reps approxLL diffs: (0.00,0.10) iter 78: time=272.08 for 3 active reps approxLL diffs: (0.02,0.06) iter 79: time=276.61 for 3 active reps approxLL diffs: (0.02,0.03) iter 80: time=274.69 for 3 active reps approxLL diffs: (0.01,0.04) iter 81: time=206.51 for 1 active reps approxLL diffs: (0.06,0.06) iter 82: time=209.50 for 1 active reps approxLL diffs: (0.09,0.09) iter 83: time=205.97 for 1 active reps approxLL diffs: (0.09,0.09) iter 84: time=214.66 for 1 active reps approxLL diffs: (0.07,0.07) iter 85: time=205.77 for 1 active reps approxLL diffs: (0.10,0.10) iter 86: time=208.05 for 1 active reps approxLL diffs: (0.26,0.26) iter 87: time=206.38 for 1 active reps approxLL diffs: (0.50,0.50) iter 88: time=203.98 for 1 active reps approxLL diffs: (0.44,0.44) iter 89: time=203.53 for 1 active reps approxLL diffs: (0.23,0.23) iter 90: time=207.86 for 1 active reps approxLL diffs: (0.13,0.13) iter 91: time=205.83 for 1 active reps approxLL diffs: (0.12,0.12) iter 92: time=204.52 for 1 active reps approxLL diffs: (0.16,0.16) iter 93: time=205.29 for 1 active reps approxLL diffs: (0.17,0.17) iter 94: time=213.19 for 1 active reps approxLL diffs: (0.09,0.09) iter 95: time=209.53 for 1 active reps approxLL diffs: (0.02,0.02) iter 96: time=205.18 for 1 active reps approxLL diffs: (0.01,0.01) Converged at iter 96: approxLL diffs each have been < LLtol=0.01 Time breakdown: dgemm = 78.1%, memory/overhead = 21.9% Filtering to SNPs with chisq stats, LD Scores, and MAF > 0.01 # of SNPs passing filters before outlier removal: 579570/705881 Masking windows around outlier snps (chisq > 437.5) # of SNPs remaining after outlier window removal: 572077/579570 Intercept of LD Score regression for ref stats: 1.157 (0.017) Estimated attenuation: 0.145 (0.016) Intercept of LD Score regression for cur stats: 1.151 (0.017) Calibration factor (ref/cur) to multiply by: 1.005 (0.002) Time for computing Bayesian mixed model assoc stats = 40735.6 sec === Computing and writing betas for risk prediction === Beginning variational Bayes iter 1: time=450.08 for 1 active reps iter 2: time=210.62 for 1 active reps approxLL diffs: (17337.47,17337.47) iter 3: time=212.46 for 1 active reps approxLL diffs: (3247.15,3247.15) iter 4: time=207.65 for 1 active reps approxLL diffs: (1189.32,1189.32) iter 5: time=208.87 for 1 active reps approxLL diffs: (589.75,589.75) iter 6: time=211.73 for 1 active reps approxLL diffs: (358.38,358.38) iter 7: time=207.78 for 1 active reps approxLL diffs: (228.04,228.04) iter 8: time=207.33 for 1 active reps approxLL diffs: (147.33,147.33) iter 9: time=206.08 for 1 active reps approxLL diffs: (93.54,93.54) iter 10: time=208.71 for 1 active reps approxLL diffs: (68.41,68.41) iter 11: time=206.77 for 1 active reps approxLL diffs: (48.71,48.71) iter 12: time=210.71 for 1 active reps approxLL diffs: (34.39,34.39) iter 13: time=209.33 for 1 active reps approxLL diffs: (27.91,27.91) iter 14: time=211.40 for 1 active reps approxLL diffs: (24.84,24.84) iter 15: time=210.82 for 1 active reps approxLL diffs: (22.76,22.76) iter 16: time=211.34 for 1 active reps approxLL diffs: (21.90,21.90) iter 17: time=212.99 for 1 active reps approxLL diffs: (13.03,13.03) iter 18: time=214.63 for 1 active reps approxLL diffs: (7.15,7.15) iter 19: time=210.42 for 1 active reps approxLL diffs: (4.87,4.87) iter 20: time=209.57 for 1 active reps approxLL diffs: (3.56,3.56) iter 21: time=204.60 for 1 active reps approxLL diffs: (4.19,4.19) iter 22: time=208.96 for 1 active reps approxLL diffs: (5.75,5.75) iter 23: time=211.74 for 1 active reps approxLL diffs: (6.01,6.01) iter 24: time=205.19 for 1 active reps approxLL diffs: (5.07,5.07) iter 25: time=209.15 for 1 active reps approxLL diffs: (4.87,4.87) iter 26: time=212.05 for 1 active reps approxLL diffs: (5.27,5.27) iter 27: time=209.98 for 1 active reps approxLL diffs: (5.67,5.67) iter 28: time=214.36 for 1 active reps approxLL diffs: (4.03,4.03) iter 29: time=211.24 for 1 active reps approxLL diffs: (3.06,3.06) iter 30: time=207.50 for 1 active reps approxLL diffs: (4.82,4.82) iter 31: time=211.42 for 1 active reps approxLL diffs: (3.55,3.55) iter 32: time=221.10 for 1 active reps approxLL diffs: (0.62,0.62) iter 33: time=225.63 for 1 active reps approxLL diffs: (0.64,0.64) iter 34: time=224.78 for 1 active reps approxLL diffs: (1.87,1.87) iter 35: time=211.14 for 1 active reps approxLL diffs: (2.19,2.19) iter 36: time=210.35 for 1 active reps approxLL diffs: (1.74,1.74) iter 37: time=222.45 for 1 active reps approxLL diffs: (1.67,1.67) iter 38: time=219.06 for 1 active reps approxLL diffs: (1.26,1.26) iter 39: time=224.58 for 1 active reps approxLL diffs: (0.85,0.85) iter 40: time=225.23 for 1 active reps approxLL diffs: (1.41,1.41) iter 41: time=221.22 for 1 active reps approxLL diffs: (3.26,3.26) iter 42: time=219.35 for 1 active reps approxLL diffs: (1.04,1.04) iter 43: time=222.15 for 1 active reps approxLL diffs: (0.66,0.66) iter 44: time=218.87 for 1 active reps approxLL diffs: (2.47,2.47) iter 45: time=227.74 for 1 active reps approxLL diffs: (1.73,1.73) iter 46: time=219.03 for 1 active reps approxLL diffs: (0.49,0.49) iter 47: time=219.09 for 1 active reps approxLL diffs: (0.23,0.23) iter 48: time=222.84 for 1 active reps approxLL diffs: (0.09,0.09) iter 49: time=220.70 for 1 active reps approxLL diffs: (0.04,0.04) iter 50: time=218.67 for 1 active reps approxLL diffs: (0.02,0.02) iter 51: time=225.63 for 1 active reps approxLL diffs: (0.01,0.01) iter 52: time=224.18 for 1 active reps approxLL diffs: (0.01,0.01) iter 53: time=219.74 for 1 active reps approxLL diffs: (0.02,0.02) iter 54: time=222.29 for 1 active reps approxLL diffs: (0.02,0.02) iter 55: time=218.02 for 1 active reps approxLL diffs: (0.02,0.02) iter 56: time=216.56 for 1 active reps approxLL diffs: (0.02,0.02) iter 57: time=222.84 for 1 active reps approxLL diffs: (0.01,0.01) iter 58: time=222.99 for 1 active reps approxLL diffs: (0.01,0.01) Converged at iter 58: approxLL diffs each have been < LLtol=0.01 Time breakdown: dgemm = 51.5%, memory/overhead = 48.5% Time for computing and writing betas = 12716.5 sec Calibration stats: mean and lambdaGC (over SNPs used in GRM) (note that both should be >1 because of polygenicity) Mean LINREG: 2.21949 (705850 good SNPs) lambdaGC: 1.49424 Mean BOLT_LMM_INF: 2.34655 (705850 good SNPs) lambdaGC: 1.50801 Mean BOLT_LMM: 2.40324 (705850 good SNPs) lambdaGC: 1.52086 Note that LINREG may be confounded by a factor of 1.00012 === Streaming genotypes to compute and write assoc stats at all SNPs === Time for streaming genotypes and writing output = 7443.01 sec Total elapsed time for analysis = 154305 sec