Preparing and Validating Molecular Data
Molecular data files are submitted in conjunction with necessary, descriptive data about molecular files and samples, called metadata. The DCC requires that metadata is submitted at the same time as the molecular data to facilitate automated downstream analysis. Metadata is submitted in JSON format, called a payload. The payload is validated against a JSON Schema to ensure data quality. This metadata is submitted to the ARGO metadata repository, called Song. To prepare the molecular metadata, you will need to follow these steps:
- Prepare metadata payloads according to the correct format (described below).
- Validate the metadata payload(s) and molecular files with the ARGO
seq-toolsclient. - Once the metadata has passed validation, both the metadata payload and the molecular data files can be submitted together.
Preparing Metadata Payloads
Understanding the Payload Format
The first step of submitting molecular data is to prepare the metadata payloads conforming to the most recent JSON Schema that has been defined by the DCC. You can find the most recent schema using this request:
The main types of submission are for Tumour and Normal samples; each of these types has a specific set of data rules to follow in the metadata format. The following are examples of correctly formatted payloads for both Tumour and Normal sample types:
- Normal Sample Metadata
- Tumour Sample Metadata
See github for the same examples of a correctly formatted payload from a normal sample and a tumour sample.
Depending on the type of sequencing that was done, the read_groups section of the payload will need to be adjusted. The following are examples of correctly formatted read_groups that meet the rules for paired-end vs single-end sequencing:
- BAM (paired)
- BAM (single)
- FASTQ (paired)
- FASTQ (single)
File and Data Validation Rules
Sequencing data of both BAM and FASTQ type files are accepted, with the exception of interleaved FASTQ files. Metadata payloads like the example above must follow these data validations:
- Compression of FASTQ files is required; both gzip (suffix .fq.gz or .fastq.gz) or bz2 (suffix .fq.bz2 or .fastq.bz2) are supported.
- For both FASTQ and BAM submissions, all files in the
filessection must be unique in the payload. - All
read_groupsin the payload and BAM header must belong to a single sample. - All
submitter_read_group_idmust be unique across the payload. read_group_id_in_bammust be unique within a single BAM.platform_unitsmust be unique across the payload with a one-to-one relationship withsubmitter_read_group_id.- The total number of
read_groupobjects must match the number specified inread_group_count. - For FASTQ paired-end sequencing, both
file_r1andfile_r2are required.file_r1andfile_r2must not be the same file. - For BAM paired-end sequencing, both
file_r1andfile_r2are required. These must be the same file. - For single-end sequencing, only
file_r1is required (file_r2must be populated withnull). - For FASTQ submission, no file can appear more than once in
file_r1orfile_r2across read group objects.
Understanding the Payload Fields
note
indicates a field that must be included in the metadata payload or it will immediately fail submission.
The metadata payload is broken down into 4 sections: experiment, samples, read groups and files, with all other fields being at the root level. Each section must be submitted in the payload.
The samples section contains details of the clinical data and key sample descriptors related to the submitted files. In order to submit a metadata payload, samples must be registered in the ARGO Data Platform. For permissible values of all fields in the samples section, please see the Sample Registration file of the Data Dictionary. If the data for a sample is different than what has been registered, metadata submission will fail.
A description of all other payload sections is shown below:
| Payload Field | Payload Section | Attribute | Description | Permissible Values |
|---|---|---|---|---|
| studyId | | ARGO Program ID, the unique identifier of your program. If you have logged into the ARGO Data Platform, this is the Program ID that you see in the Program Services area. | ||
| analysisType | | The type of molecular data that is being submitted in the payload. | sequencing_experiment | |
| submitter_sequencing_ experiment_id | experiment | | The unique identifier of the sequencing experiment. | String values that meet the regular expression ^[a-zA-Z0-9]{1}[a-zA-Z0-9\\-_\\.:']{0,98}[a-zA-Z0-9]{1}$ or null. |
| platform | experiment | | The sequencing platform type used in data generation. | CAPILLARY, LS454, ILLUMINA, SOLID, HELICOS, IONTORRENT, ONT, PACBIO, Nanopore, BGI |
| platform_model | experiment | The model number of the sequencing machine used in data generation. | Any string value or null. | |
| experimental_strategy | experiment | | The primary experimental method. For sequencing data it refers to how the sequencing library was made. | WGS, WXS, RNA-Seq, Bisulfite-Seq |
| sequencing_date | experiment | Date sequencing was performed. | datetime format, for example: 2019-06-16 or 2019-06-16T20:20:39+00:00 | |
| read_group_count | | The number of read groups in the molecular files being submitted. | A minimum of 1 is required. | |
| read_group_id_in_bam | read_groups | Optional field indicating the @RD ID in the BAM. If submitted, this will be used to map the @RG ID in the BAM header to the submitter_read_group_id in the payload. After submission, the @RG ID in the payload will be used for all future headers. This cannot be submitted for FASTQ files. | String value must meet the regular expression ^[a-zA-Z0-9\\-_:\\.']+$. | |
| submitter_read_ group_id | read_groups | | The identifier of a read group; must be unique within each payload. | String values that meet the regular expression ^[a-zA-Z0-9\\-_:\\.']+$ or null. |
| platform_unit | read_groups | | Unique identifier including the {FLOWCELL_BARCODE}.{LANE}.{SAMPLE_BARCODE}. The {FLOWCELL_BARCODE} refers to the unique identifier for a particular flow cell. The {LANE} indicates the lane of the flow cell and the {SAMPLE_BARCODE} is a sample/library-specific identifier. For non-multiplex sequencing, platform unit and read group have a one-to-one relationship. | Any string value. |
| is_paired_end | read_groups | | Indicate if paired-end sequencing was performed. | true, false |
| file_r1 | read_groups | | Name of the sequencing file containing reads from the first end of a sequencing run. | Any string value. Must match a fileName identified in the files section. |
| file_r2 | read_groups | | Name of the sequencing file containing reads from the second end of a paired-end sequencing run. Required if and only if paired-end sequencing was done. | Any string value or null. Must match a fileName identified in the files section. |
| read_length_r1 | read_groups | | Length of sequencing reads in file_r1; this corresponds to the number of sequencing cycles of the first end. | Integer with a minimum value of 20 or null. |
| read_length_r2 | read_groups | | Length of sequencing reads in file_r2; this corresponds to the number of sequencing cycles of the second end. | Integer with a minimum value of 20 or null. |
| insert_size | read_groups | | For paired-end sequencing, the average size of sequences between two sequencing ends. Required only for paired-end sequencing. | Integer with a minimum value of 0 or null. |
| sample_barcode | read_groups | | According to the SAM specification, this is the expected barcode bases as read by the sequencing machine in the absence of errors. | Any string value or null. |
| library_name | read_groups | | Name of a sequencing library made from a molecular sample or a sample pool (multiplex sequencing). | Any string value. |
| fileName | files | | Name of the file. | String values must meet the regular expression ^[A-Za-z0-9_\\.\\-\\[\\]\\(\\)]+$. No paths are allowed in the file name. |
| fileSize | files | | Size of the file, in bytes. | |
| fileMd5sum | files | | Computed md5sum of the file. This must match the payload, and the md5sum computed when the file is uploaded. | String values must meet the regular expression ^[a-fA-F0-9]{32}$ |
| fileType | files | | Data format of sequencing files. | BAM, FASTQ |
| fileAccess | files | | The permission level of a file for public access. | open, controlled |
| dataType | files | | Descriptor of the type of file being submitted. | Submitted Reads |
Validating Metadata Payloads
It is very important that molecular data is submitted with valid metadata. As a helpful tool during metadata preparation, we have prepared a validation client that can be run on your data locally before officially submitting.
Validation will help you ensure your data is formatted correctly (with accurate identifier assignments between metadata and molecular data files) and that your submission goes smoothly. It also helps the DCC ensure that the downstream Analysis Pipelines will function seamlessly.
This validation client will check:
This validation client will check:
- Metadata in the payload aligns with the BAM/FASTQ files.
- Each submission is for a single sample.
- Each payload is formatted according to the data validation rules stated above.
Installing the Validation Client
- Client
- Docker
A pre-requisite is to have the samtools package installed, which is used to retrieve BAM header information. To read JSON/JSONL outputs in a friendly manner, we also suggest installing the jq package.
Once samtools is installed, clone the repository:
Using pip, install seq-tools:
You can verify that the installation worked by checking the version:
To update the client to the latest version, pull the latest version of the repository and run:
Structured Directory Validation
One method to quickly complete validation is to structure your data in a predetermined format that seq-tools is expecting. Under a submission directory (e.g. molecular-data-submission in the following example), compile each submission as a folder, with the JSON payload and the associated data files in the same folder.
Each submission folder must contain:
- The metadata payload as a
.jsonfile. Only one.jsonfile must be present in the submission folder. - All molecular data files associated to that payload, or symlinks to their location.
For example:
Once your data is ready, make sure you are under the parent folder of the submission directory and use seq-tools to validate all of the data:
- Client
- Docker
You will see an interactive prompt of the results of validation of each metadata payload as the tool runs. Once completed, view the details of the validation reports under the current folder:
Scriptable Location Validation
If data files are located in a different directory than where the metadata payloads are listed, and you do NOT want to reorganize them into the above structured directory, you can use the '-d' option to provide the data files location. You can also use this to validate all metadata payloads in a directory at once using a wildcard.
For example:
- Client
- Docker
You can use these methods to build your own script to validate payloads in different locations.
Understanding the Validation Report
Validation report summaries can be found in the folder where you ran seq-tools. Reports are separated into three categories:
PASS: This status indicates that the payload(s) are ready for submission.
PASS-with-WARNING: This status indicates that the payload(s) are ready for submission, however there may be a parameter you want to double check before submission.
INVALID: This status indicates that the payload(s) are not ready for submission.
PASS: This status indicates that these payload(s) are ready for submission.
PASS-with-WARNING: This status indicates that these payload(s) are ready for submission, however there may be a parameter you want to double check before submission.
INVALID: This status indicates that these payload(s) are not ready for submission.
In the INVALID summary report, each reason for failure is listed. Errors must be fixed before you attempt to submit this payload. To view all reports that are generated: