MonoRes - Local resolution

MonoRes is an algorithm to estimate the local resolution in cryoEM maps. It estimates the local resolution by establishing an statistical comparison at different resolutions between the local signal (obtained from the monogenic amplitude) and the noise. The limit at which the local signal cannot be distisguished from noise will be the local resolution of the region.

Reference: This method can be detailed read in:

J.L. Vilas et al., MonoRes: Automatic and Accurate Estimation of Local Resolution for Electron Microscopy Maps, Structure 26, 337-344 (2018).

The understading of local resolution can be misleading. Local resolution should be understood as a relative measurement between different regions. Therefore it is not an absolute measurement. For a discussion about the good practices in local resolution, the developers of local resolution wrote an article to inform about their use, see

J.L. Vilas et al., Local resolution estimates of cryoEM reconstructions, Current opinion in structural biology 64, 74-78 (2020).

Install

Using conda/mamba (recommended, environment.yml included in this folder):

conda env create -f environment.yml
conda activate monores

By default environment.yml installs the CPU-only PyTorch build. To use a GPU, edit the file: comment out the pytorch / cpuonly pair under the CPU section and uncomment the matching pytorch / pytorch-cuda=... block for your installed NVIDIA driver, then re-run conda env create.

Or with plain pip:

pip install --user torch mrcfile numpy
# CPU-only torch (much smaller):
# pip install --user --index-url https://download.pytorch.org/whl/cpu torch

Usage

Before each execution the conda environment must be activated

# To activate conda enviroinment
conda activate monores

Then, MonoRes can be executed

# two half maps (recommended)
python3 monores.py \
    --vol half1.mrc --vol2 half2.mrc --mask mask.mrc \
    --sampling_rate 1.34 --maxRes 25 --step 0.25 \
    -o monores_out

# single map
python3 monores.py --vol map.mrc --mask mask.mrc \
    --sampling_rate 1.34 --maxRes 25 -o monores_out

Other options can be checked with

# To check all monores options
python3 monores.py --help

Watch out for --minRes / --maxRes

This is upstream behaviour, faithfully reproduced:

• --minRes (default 30) is overwritten with 2 * sampling_rate on the first line of run(), so the CLI value has no effect at all.
• The sweep starts at resolution = maxRes and decreases by --step until it drops below Nyquist. --maxRes is therefore the low-frequency start of the range, despite its name.

Outputs

Written into the -o folder as MRC mode 2 (float32), matching Xmipp (ImageBase::writeMRC maps both Image<double> and Image<int> to mode 2):

File	Content
meanMap.mrc	0.5*(V1+V2) -- half-map mode only
refinedMask.mrc	binary mask of voxels that got a resolution value
monoresResolutionChimera.mrc	smoothed resolution map over the whole box (for visualisation)
monoresResolutionMap.mrc	smoothed resolution map, zero outside the mask