feather_files/README.md

In CRI-iAtlas/iatlas-data: iAtlas Data

iAtlas Data Structures

When importing data into iAtlas, it is very important that the following conventions are followed. Doing so will get the new data into the iAtlas database and make it available for the app.

Data Model

Information on the data model can be found in the data_model folder which contains this README.md file.

File Format

• feather files

All data should come into the iAtlas application in the form of feather files. Feather files allow for fast reading and help ensure structural integrity.

Data Locations

All data (feather files) should be located in the iAtlas Synapse directory

Within the feather_file directory, data files should be segregated into folders as follows:

• copy_number_results
• datasets
• driver_results
• edges
• features
• gene_types
• genes
• mutation_codes
• mutation_types
• mutations
• nodes
• patients
• publications
• relationships
• datasets_to_tags
• features_to_samples
• genes_to_samples
• genes_to_types
• samples_to_mutations
• samples_to_tags
• tags_to_tags
• samples
• slides
• tags

Feather File Structure

Data files in each folder MUST follow a specific convention for that data type. The files can be named as is deemed most descriptive and MUST end in .feather.

Column names MUST be spelled exactly as shown in this document.

The conventions for the feather files in each folder are as follows:

• copy_number_results

#### Copy Number Results Column Names

• feature

  The name of a feature. These unique names MUST exist in data in the features folder.

  type - (character)

• entrez

  The entrez id of a gene. These genes MUST exist in data in the genes folder.

  type - (numeric)

• dataset

  The name of a dataset. These unique names MUST exist in data in the datasets folder.

  type - (character)

• tag

  The tag name associated with this copy number result. These tags MUST exist in data in the tags folder.

  type - (character)

• direction

  The direction of this copy number result.

  type - (DIRECTION_ENUM)

• mean_normal

  The mean normal value this copy number result.

  type - (numeric)

• mean_cnv

  The mean normal cnv this copy number result.

  type - (numeric)

• p_value

  The p value associated with this copy number result.

  type - (numeric)

• log10_p_value

  The log10 p value associated with this copy number result.

  type - (numeric)

• t_stat

  The t stat value of this copy number result.

  type - (numeric)

### datasets

#### Datasets Column Names

• name

  The name of a the dataset. Must be unique, must not use any charcaters besides letters, number and underscores.

  type - (character)

• display

  A display name for the dataset.

  type - (character)

### driver_results

#### Driver Results Column Names

• feature

  The name of a feature. These unique names MUST exist in data in the features folder.

  type - (character)

• dataset

  The name of a dataset. These unique names MUST exist in data in the datasets folder.

  type - (character)

• entrez

  The entrez id of a gene. These genes MUST exist in data in the genes folder.

  type - (numeric)

• mutation_code

  The mutation code associated with this driver result. These mutation codes MUST exist in data in the mutation_codes folder.

  type - (character)

• tag

  The tag name associated with this driver result. These tags MUST exist in data in the tags folder.

  type - (character)

• p_value

  The p value associated with this driver result.

  type - (numeric)

• fold_change

  The fold change value associated with this driver result.

  type - (numeric)

• log10_p_value

  The log10 p value associated with this driver result.

  type - (numeric)

• log10_fold_change

  The log10 fold change value associated with this driver result.

  type - (numeric)

• n_wt

  The number of "Wild Type" genes associated with this driver result.

  type - (integer)

• n_mut

  The number of "Mutant" genes associated with this driver result.

  type - (integer)

• edges

#### Edges Column Names

• name

  The name of the edge. Each edge must have a unique name, we suggest <dataset_name>_<network_name>_<row_number>

  type - (character)

• node1

  The node the edge is starting from. This must be in the name column of the corresponding nodes file.

  type - (integer)

• node2

  The node the edge is ending at. This must be in the name column of the corresponding nodes file.

  type - (integer)

• label (optional)

  The label of the edge.

  type - (character)

• score (optional)

  The numeric value of the edge.

  type - (numeric)

• features

#### Features Column Names

• name

  The name of the feature.

  type - (character)

• display

  A friendly display name for the feature.

  type - (character)

• class

  The class of the feature. If the feature does not have a class, use Miscellaneous.

  type - (character)

• method_tag (optional)

  The method tag of the feature.

  type - (character)

• order (optional)

  The prefered order of priority for the feature.

  type - (integer)

• unit (optional)

  The unit used for the value of the feature.

  type - (UNIT_ENUM)

• gene_types

#### Gene Type Column Names

• name

  The name of the gene type.

  type - (character)

• display

  A friendly display name for the gene type.

  type - (character)

• genes

#### Gene Column Names

• entrez (required)

  The entrez identifier of the gene. This is used through out the app to uniquely identify the gene. This is REQUIRED.

  type - (numeric)

• hgnc

  The Hugo Id of the gene.

  type - (character)

• description

  A description of the gene.

  type - (character)

• friendly_name (optional)

  A human friendly display name for the gene.

  type - (character)

• io_landscape_name (optional)

  The IO Landscape name for the gene.

  type - (character)

• gene_family (optional)

  The gene family of the gene.

  type - (character)

• gene_function (optional)

  The gene function of the gene.

  type - (character)

• immune_checkpoint (optional)

  The immune checkpoint for the gene.

  type - (character)

• pathway (optional)

  The pathway of the gene.

  type - (character)

• super_category (optional)

  The super category of the gene.

  type - (character)

• therapy_type (optional)

  The therapy type of the gene.

  type - (character)

• mutation_codes

#### Mutation Code Column Names

• code

  The mutation code. This is REQUIRED.

  type - (character)

• mutation_types

#### Mutation Type Column Names

• name

  The name of the mutation type.

  type - (character)

• display

  A friendly display name for the mutation type.

  type - (character)

• mutations

#### Mutation Column Names

• entrez (required)

  The entrez id of a gene. These genes MUST exist in data in the genes folder.

  type - (numeric)

• mutation_code (required)

  The code (name) of a mutation code. These mutation codes MUST exist in data in the mutation_codes folder.

  type - (character)

• mutation_type

  The name of a mutation type. These mutation types MUST exist in data in the mutation_types folder.

  type - (character)

• nodes

#### Node Column Names

A node may use a gene OR a feature. One of these is REQUIRED.

• name

  The name of the nodes. Each node must have a unique name, we suggest <dataset_name>_<network_name>_<row_number>

  type - (character)

• entrez

  The entrez id of a gene. These genes MUST exist in data in the genes folder.

  type - (numeric)

• feature

  The name of the feature. These features MUST exist in data in the features folder.

  type - (character)

• network

  The network tag related to the node. These tags MUST exist in data in the tags folder.\

  type - (character)

• dataset

  The name of a dataset. These unique names MUST exist in data in the datasets folder.

  type - (character)

• tag

  The a tag related to the node. These tags MUST exist in data in the tags folder.\ For a node to be used in edges, this column name MUST also exist in the edges data.

  type - (character)

• tag_\ (optional)

  Additional tags related to the node. These tags MUST exist in data in the tags folder.\ The column name MUST start with tag but may be followed by a dot (_) and some additional descriptive text. ie tag_second or tag_01. There may be as many tag columns as needed.\ This column name MUST also exist in the nodes data.

  type - (character)

• label (optional)

  The label of the edge.

  type - (character)

• score (optional)

  The numeric value of the node.

  type - (numeric)

• patients

#### Patients Column Names

• barcode

  The unique identifier representing a patient.

  type - (character)

• age (optional)

  The age of the patient.

  type - (character)

• ethinicity (optional)

  The ethinicity of the patient.

  type - (character)

• gender (optional)

  The gender of the patient.

  type - (character)

• height (optional)

  The height of the patient.

  type - (character)

• race (optional)

  The race of the patient.

  type - (character)

• weight (optional)

  The weight of the patient.

  type - (character)

• samples

#### Sample Column Names

• name

  The unique identifier representing the sample.

  type - (character)

• patient_barcode

  The unique identifier representing a patient related to the sample. The patient MUST exist in the data in the patients folder.

  type - (character)

• dataset

  The name of a the dataset. These unique names MUST exist in data in the datasets folder.

  type - (character)

• slides

#### Slide Column Names

• name

  The unique identifier representing the slide.

  type - (character)

• patient_barcode

  The unique identifier representing a patient related to the slide. The patient MUST exist in the data in the patients folder.

  type - (character)

• tags

#### Tag Column Names

Tags may be used to group various pieces of data. At a base level, a tag is simply a string (with some descriptive meta data). Multpile pieces of data may be related by tagging them. Tags may even be tagged to create the semblance of hierarchy.

• name

  The unique identifying name of the tag.

  type - (character)

• characteristics

  Any identifying characteristics of the tag.

  type - (character)

• display

  A human friendy display name for the tag.

  type - (character)

• color

  A specific hex value to represent the tag by color.

  type - (character)

• publications

#### Publications Column Names

• pubmed_id

  The unique id at "https://pubmed.ncbi.nlm.nih.gov/{id}"

  type - (integer)

• journal

  The journal published in

  type - (character)

• first_author_last_name

  The last name of the first author

  type - (character)

• year

  The year published

  type - (integer)

• title

  The name of the publication

  type - (character)

• relationships

Often data is about relationships. The following folders are for data relationships. Each relationship depends on the original dat pieces being represented in their respective folders.

• publications_to_genes

  Publications To Genes Column Names

  • pubmed_id

  The pubmed id of the publication. These unique ids MUST exist in data in the publications folder.

  type - (integer)

  • entrez

  The entrez id of the gene. These unique ids MUST exist in data in the genes folder.

  type - (integer)

• datasets_to_tags

  Datasets To Tags Column Names

  • dataset

  The name of a the dataset. These unique names MUST exist in data in the datasets folder.

  type - (character)

  • tag

  The name of the tag. These unique names MUST exist in data in the tags folder.

  type - (character)

• features_to_samples

  Feature to Sample Column Names

  • feature

  The name of the feature. These features MUST exist in data in the features folder.

  type - (character)

  • sample

  The name of the sample. These samples MUST exist in data in the samples folder.

  type - (character)

  • value

  The numeric value of the feature to sample relationship. The unit of the value is expressed in the features data.

  type - (numeric)

• genes_to_samples

  Gene to Sample Column Names

  • entrez

  The entrez id of a gene. These genes MUST exist in data in the genes folder.

  type - (numeric)

  • sample

  The name of the sample. These samples MUST exist in data in the samples folder.

  type - (character)

  • rna_seq_expr

  The unique numeric RNA sequence expression of the relationship between the gene and the sample.

  type - (numeric)

• genes_to_types

  Gene to Gene Type Column Names

  • entrez

  The entrez id of a gene. These genes MUST exist in data in the genes folder.

  type - (numeric)

  • gene_type

  The type of gene this specific gene is related to. These gene types MUST exist in data in the gene_types folder.

  type - (character)

• samples_to_mutations

  Sample to Mutation Code Column Names

  • sample (required)

  The name of the sample. These samples MUST exist in data in the samples folder.

  type - (character)

  • entrez (required)

  The entrez id of a gene. These genes MUST exist in data in the genes folder.

  type - (numeric)

  • mutation_code (optional - may be NA)

  The code (name) of the mutation code. These mutation codes MUST exist in data in the mutation_codes folder.

  type - (character)

  • mutation_type (optional - may be NA)

  The name of the mutation type. These mutation types MUST exist in data in the mutation_types folder.

  type - (character)

  • status

  The status of the gene in this psecific relationship. My be Wt (Wild Type) or Mut (Mutant).

  type - (STATUS_ENUM)

• samples_to_tags

  Sample to Tag Column Names

  • sample

  The name of the sample. These samples MUST exist in data in the samples folder.

  type - (character)

  • tag

  The tag related to the sample. These tags MUST exist in data in the tags folder.

  type - (character)

• tags_to_tags

  Tag to Tag Column Names

  • tag

  The name of the tag. These tags MUST exist in data in the tags folder.

  type - (character)

  • related_tag

  The tag related to the initial tag. These tags MUST exist in data in the tags folder.

  type - (character)

CRI-iAtlas/iatlas-data documentation built on July 7, 2020, 2:18 a.m.