The STAR format defines a limited hierarchy, so that expressing the micrograph parameter hierarchy requires some specialized conventions. Each data block corresponds to a micrograph, and particle parameters are encoded in looped variables tightly linked to the original micrograph. Micrographs are identified by an "id" tag, and the fields-of-view they belong are identified by a "field_id" tag. Each particle is identified by a unique number, starting from 1 for the first particle in the micrograph:

Parameters for the micrograph proper are encoded as tags with single values, such as the micrograph file name, the acceleration voltage and CTF parameters. Parameters for the individual particle images picked from a micrograph are encoded in a table-like construct starting with the keyword "loop_" followed by the list of tags and columns of values. The first column should always be the particle number starting with the first particle as one and referencing all the particles in the picked particle file. Any selections of particles should be done using a selection column (not by deleting and rearranging the particle records). The loop must be followed by an empty line to indicate its end. The start of a STAR file (before any "data_" keywords) is used to record the history of the file. Every program writing a STAR file should add comments to the end of this section giving the date and time and the command line for execution. Additional comments can be inserted at any point.