EMAN2 vs Relion perceived resolution - mystery solved, and a new EMAN option

[Steve Ludtke]

There has been a perception for the last couple of years that if you solve a structure in EMAN2 and a structure in Relion that even if you get the same resolution, people feel the Relion map "looks better". I have observed the effect myself, and particularly for alpha helices at better than ~5 Å resolution, there has been a clear effect. This was very confusing because Sjors always claimed that Relion was performing a final Wiener filter based on the 'gold standard' FSC curve, combined with B-factor correction. This is exactly what EMAN2.1 does as well, so the discrepancy was difficult to explain. 

Thanks to some time Pawel spent reading the actual code, it appears that this is not what the final Relion output contained at all!  If you run the beautification program as recommended by Sjors, it does the B-factor correction, then performs a so-called Tophat filter at the 0.143 FSC resolution, that is, a sharp cutoff to zero. This means that none of the noise-reducing Wiener filtration is taking place at resolutions where the FSC is falling, and that ringing due to the sharp cutoff will enhance high resolution features.

Lo and behold, when we switch the final filter in EMAN2.1 from a Wiener filter (which in a least-squares sense is still the 'correct' filter to apply) to a Tophat filter, the maps look much sharper. Alpha-helices develop much stronger pitch and putative sidechains stick out strongly.

While I still have somewhat mixed feelings about the appropriateness of this filter, it undeniably makes the maps look 'better'. So, there is now a new option in e2refine_easy.py:

--tophat

If you use this option, your maps will look much sharper. This will not impact resolution, just how the final maps are filtered, and the effect will be strongest beyond 5 Å resolution. Indeed, I worry a little that this might do 'bad things' at very low resolution, which is why it's an option.

For your amusement, a comparison of Yifan's first TRPV1 structure and the EMAN reconstruction from the same data (this is not from their newer structure):

----------------------------------------------------------------------------
Steven Ludtke, Ph.D.
Professor, Dept. of Biochemistry and Mol. Biol.                Those who do
Co-Director National Center For Macromolecular Imaging            ARE
Baylor College of Medicine                                     The converse
slu...@bcm.edu  -or-  ste...@alumni.caltech.edu               also applies
http://ncmi.bcm.edu/~stevel