rk.score

Score MSA subsamples against experimental likelihood

rk.score calculates the experimental likelihood (log-likelihood gain, LLG) for single-shot predictions made using different MSA subsamples

Note: check out rk.msacluster to generate different MSA subsamples in the first place.

options

-h, --help                         show this help message and exit

-i I                               prefix for MSAs to use (e.g. submsa_), 
                                   working path will prepend

-o O                               name of output directory to write scoring 
                                   results to, working path will prepend

--domain_segs DOMAIN_SEGS [...]    list of residue indices for domain segmentation 
                                   (e.g. --domain_segs 50 150)

--additional_chain                 flag indicating an additional chain is present in ASU

--init_recycling INIT_RECYCLING    number of AlphaFold recycling iterations 
                                   (default: 4)

--free_flag FREE_FLAG              column name of free flag in MTZ file 
                                   (default: R-free-flags)

--testset_value TESTSET_VALUE      label value for test set reflections 
                                   (default: 1)

--voxel_spacing VOXEL_SPACING      voxel spacing used in solvent estimation 
                                   (default: 4.5)

--min_resolution MIN_RESOLUTION    minimum resolution cutoff in Ångstroms 
                                   (default: 1.0)

--chimera_profile                  enable blending full MSA profile into subset 
                                   profile (chimera-style)

--score_fullmsa                    score the full alignment in alignments/ directory

Typical command example:

rk.score <path> <file_id> -i msa_clusters/ -o msa_scores --score_fullmsa