Table 2 Statistics of performance and suggested sequences from NucleicNet and RNAcompete (RNAC)

From: A deep learning framework to predict binding preference of RNA constituents on protein surface

Figure 4 a b c d e f g h
Gene name PABPC1 PCBP2 PTBP1 RBFOX1 SNRPA SRSF2 TARDBP U2AF2
Sampled PDBID 1cvj 2py9 2adc 2err 1aud 2lec 4bs2 2g4b
RNAC ID 155 44 269 168 71 72 76 79
RNAC suggested sequence ARAAAAM CCYYCCH HYUUUYU WGCAUGM WUGCACR GGAGWD GAAUGD UUUUUYC
NucleicNet suggested sequence AAAAAAW WHCYCUWHCYCU UUUWYU URHAUGU AWUGCAH WNGAGW RURWAUGA UUDWW
PDB deposited sequence AAAAAAA AACCCUAACCCU CUCUCU UGCAUGU AUUGCAC UGGAGU GUGAAUGA UUUUU
Pearson correlation (RNAC PWM Score vs NucleicNet Score) 0.81 0.70 0.73 0.27 0.74 0.32 0.77 0.72
Welch’s t-test statistics (highest 10—lowest 10) 20.7 16 25.3 5.2 6.2 7 1.7 20.2
Welch’s t-test P-value 6.10E-13 1.90E-09 6.70E-13 2.40E-04 4.90E-05 3.90E-06 1.10E-01 8.30E-09
  1. Best matching suggested sequences between RNAC and NucleicNet are underlined. R: A/G, M: A/C, Y: C/T, H: A/C/T, W: A/T, D: A/G/T, and N: A/C/G/U