I run the program twice. Once for each. CTF estimate from the unweighted and extract from the dose weighted.

From: Collaborative Computational Project in Electron cryo-Microscopy <[log in to unmask]> on behalf of Zhang Wenjuan <[log in to unmask]>
Sent: Wednesday, October 19, 2016 7:52 PM
To: [log in to unmask]
Subject: Re: [ccpem] About MotionCor2 binning

Hi Shawn,

For MotionCor2 corrected micrographs, how can we use dose weighted mrc to extract particles while using the CTF estimation of non-dose weighted mrc? Since we can only extract particles from micrographs_ctf.star file in Relion2.0, right? Thank you very much.

Best wishes.

Wenjuan

On 19 Oct 2016, at 01:09, Shawn Zheng <[log in to unmask]> wrote:

Hi Craig,

Thanks for the good suggestion. This is certainly something that is worth of some serious effort.

Best

Shawn

On Tue, Oct 18, 2016 at 4:46 PM, Craig Yoshioka <[log in to unmask]> wrote:

No, the later frames have their high-res values downweighted.

It would be awesome if an alignment program took into account the estimated drift rate of each frame and either discarded, or down-weighted, frames with more drift (maybe it does?).

On Oct 18, 2016, at 4:36 PM, Ali Khan <[log in to unmask]> wrote:

Shawn,

Is throwing away the first few initial frames common practice when dose weighting is implemented? In this case, wouldn't dose weighting down weigh those initial frames?

Cheers,

Ali

On Tue, Oct 18, 2016 at 7:20 PM, Shawn Zheng <[log in to unmask]> wrote:

Hi Tat,

I feel your questions may be of general interest, so I send to CCPEM too.

Since the local correction is done by linear interpolation that has low-pass filtering effect, I recommend to use super-res movies as input and let MotionCor2 do the binning. This allows MotionCor2 to interpolate at super-res pixels to reduce the low-pass effect on the final sums.

Yes, throwing away a few initial frames is still a common practice given the fast motion at the beginning of exposure.

Best

Shawn

On Tue, Oct 18, 2016 at 11:24 AM, Cheng, Tat <[log in to unmask]> wrote:

Hi Shawn,

I notice MotionCor2 uses super resolution movie as input, but does it matter if I choose to use bin2 movie for processing? And in general do you throw away the first few frames?

Thanks a lot.

Tat

--

********************************************************************

Ali K. Khan

M.D.-Ph.D. Candidate

Medical Scientist Training Program

Department of Molecular Physiology and Biological Physics

University of Virginia School of Medicine and

Graduate School of Arts and Sciences

Sheridan G. Synder Translational Research Building, Rm 320

480 Ray C. Hunt Drive

Charlottesville, VA 22908

Phone: 434-243-2520

Email: [log in to unmask]

********************************************************************