A silent file is a compact file format that stores information about a structure. Unlike the PDB format, the silent format can contain information for more than one structure.
There are two encodings of silent files: Protein and Binary. They differ in how they store the coordinate info. Protein silent files store the phi/psi/chi information in a text-based column layout. Binary silent files store the full xyz atom coordinates in an ASCII-encoded version of the binary IEEE 754 double value. Protein struct silent files hold the structural information as columns, but in binary files, the structural information is saved as an ASCII string. Binary Silent Struct File is very useful for compressing multiple PDBs and saving computer space. Protein Silent Struct is usually seen in the Abinitio outputs.
The output encoding can be changed using this flag:
-out:file:silent_struct_type binary
or
-out:file:silent_struct_type protein
Protein encoding
Header
SEQUENCE Structure sequences presented by one-letter amino acid code
SCORE Rosetta score terms
Body
Columns 1-4 Residue sequence number
Columns 5-7 Secondary structure one letter code
Columns 8-17 Phi angle
Columns 18-26 Psi angle
Columns 27-35 Omega angle
Columns 36-44 CA atom coordinates x
Columns 45-53 CA atom coordinates y
Columns 55-62 CA atom coordinates z
Columns 64-98 Chi angle real data if possible
Binary encoding
Header
SEQUENCE Structure sequences presented by one-letter amino acid code
SCORE Rosetta score terms
Body
Binary contents
The extract_pdbs
application can convert silent files into PDB files. Information about the options can be found in the integration test folder /Rosetta/main/tests/integration/tests/extract_pdbs/.
One can use methods native to Linux to extract specific poses from a silent file. For example:
#Getting the header of silent file
head -3 file.silent > new.silent
#Appending data from pose with tag 'result_0001' to the new silent file
grep 'result_0001' file.silent >> new.silent