Contrast transfer function correction model based on sine and Gaussian modulation and spline interpolation

Abstract

In the course of 3D reconstruction for protein molecules based on single-particle cryoelectron microscopy, CTF correction has played an important role on obtaining the high resolution molecular structure. However, the traditional CTF correction model has some problems. Firstly, the reconstruction result would be amplified incorrectly in low frequency and instable in high frequency, because CTF value is very small in the two frequency regions. Secondly, the linear interpolation would affect the precision of adjusted data, because it could not match the sinusoidal shape of CTF curve. Thus, the above two issues are studied and a new CTF correction model is proposed. On the one hand, the sine and Gaussian modulation are applied in the low-frequency and high-frequency phase of CTF curve respectively. Thus, the shape of CTF curve is enhanced. On the other hand, the cubic spline interpolation is used for computing the CTF value in the fractional Fourier radius. The accuracy of the data acquisition is improved subsequently. Thereby, the resolution of reconstruction molecular structure is increased. Finally, the reconstruction experiment is carried out with the cryo-EM images of HBV. The experimental results indicate that the structure from our model has more detailed feature than the one from the old model. The resolution of reconstructed strcuture with our method is 7.8 angstrom, which is better than 8.06 angstrom of the one with traditional CTF model.