library(prego)
load("/net/mraid20/ifs/wisdom/tanay_lab/tgdata/users/yonshap/proj/mmcortex/output/MPRA/mpra_prego_data.rda")
prego_mpra <- regress_pwm(sequences = tbl_mpra_seqs, response = tbl_mpra, motif_num = 4)
ℹ Using 5 bins of size 40 bp
ℹ Running regression for 4 motifs
── Running first regression ──
ℹ Using 5 bins of size 40 bp
ℹ Using "r2" as the final metric
ℹ Number of response variables: 8
── Generate candidate kmers
── Regress each candidate kmer
ℹ Running regression on 26 candidate kmers
• Bidirectional: TRUE
...
• kmer: "CGGTG*ACT", score (r2): 0.00160458151361125
• kmer: "GTA*GACGG", score (r2): 0.00153592844677623
• kmer: "GTAAG*CGG", score (r2): 0.00301175495505033
• kmer: "TA*GACGGT", score (r2): 0.000621375777523677
• kmer: "TAAG*CGGT", score (r2): 0.000785608254398136
• kmer: "TATAG*TAA", score (r2): 0.00923729984210879
ℹ Best match in the database: "HOCOMOCO.ATOH1_HUMAN.H11MO.0.B", cor: 0.855
ℹ Best motif: "***AGATA*GAT***", score (r2): 0.0295539826452405
ℹ R2 for models 1, 2, 3, and 4: 1, 0.0118910457065785, 0.00315125565289002, 0.0129680234050577, 0.00660287176430303, 0.000551345644021713, 0.00283192566909634, 0.00212289201685585, 0.0118910457065785, 1, 9.46374496342389e-05, 0.00531800308025085, 0.000594129157913513, 0.000226539758160116, 5.15332172466543e-05, 0.000432008692663793, 0.00315125565289002, 9.4637449634239e-05, …, 0.00957589407994158, and 1
ℹ Improvement in R2: 0.988108954293421, 0, -0.00873979005368847, 0.00107697769847918, -0.00528817394227547, -0.0113397000625568, -0.00905912003748216, -0.00976815368972265, 6.93889390390723e-18, 0.988108954293421, -0.0117964082569443, -0.00657304262632764, -0.011296916548665, -0.0116645059484184, -0.0118395124893318, -0.0114590370139147, -0.00873979005368848, -0.0117964082569443, …, -0.00231515162663691, and 0.988108954293421
Error in data.frame(model = 1:motif_num, score = scores, comb_score = comb_scores, :
arguments imply differing number of rows: 4, 25, 193
In addition: Warning message:
In comb_scores[1] <- res$r2 :
number of items to replace is not a multiple of replacement length