The following attributes have disappeared from experiment objects:
system_slims
organ_slims
biosample_term_name
developmental_slims
biosample_type
biosample_synonyms
biosample_term_id
The following columns are new:
experiment_classification
biosample_ontology
biosample_type and biosample_term_name were part of the final encode_df.
biosample_type entries are indices into the new biosample_type table (EX: "/biosample-types/cell_line_EFO_0001182/", "/biosample-types/primary_cell_CL_1000458/").
The biosample_type table has the following columns. Columns which directly reference old experiments columns are marked with *. Columns which were part of encode_df are marked with **
dbxrefs
status
developmental_slims *
cell_slims
uuid
organ_slims *
type **
system_slims *
synonyms *
schema_version
classification
name
term_id *
id
term_name **
The lack of those columns cause export_ENCODEdb_matrix to fail when reordering columns. Three possible actions must be taken:
At the very least, the function mustn't crash anymore and drop the missing columns silently.
If possible, the columns from biosample_type which were previously in encode_df should now be joined into the experiment table so that information isn't lost.
To go above and beyond the call of duty, the new columns of biosample_type which have no equivalent in the old data frame should also be added.
The following attributes have disappeared from experiment objects:
The following columns are new:
biosample_type and biosample_term_name were part of the final encode_df.
biosample_type entries are indices into the new biosample_type table (EX: "/biosample-types/cell_line_EFO_0001182/", "/biosample-types/primary_cell_CL_1000458/").
The biosample_type table has the following columns. Columns which directly reference old experiments columns are marked with *. Columns which were part of encode_df are marked with **
The lack of those columns cause export_ENCODEdb_matrix to fail when reordering columns. Three possible actions must be taken: