Filter Mode
Compared with single - nucleotide polymorphism (SNP), the filtering and screening of SV often lack a unified standard, and the screening attributes are more diverse (possibly including INFO, QUAL, GT fields, etc.).
To perform more comprehensive SV filtering, SDFA has a rich set of built - in filtering functions. Specifically, it mainly covers two aspects of filtering:
Genotype level：Set quality control information for genotype filtering - set those that do not meet the conditions as./.SV level：Set a custom function to filter multiple VCF fields such asCHRandIDof the current SV record
The specific instructions are
java -jar sdfa.jar filter -d [input_dir] -o [output_dir] [options]
Genotype Filter
Use the following instructions for filtering:
--filter-gty <format_attr> <function>
The above command line filters the specified format_attr, and function is an expression that returns a boolean value.
To facilitate users to directly operate on the values of the quality control attributes of each genotype, we directly use value in the function to represent the value of a single genotype. Now, let's take an example:
[!NOTE|label:Example 1]
We filter the genotypes of all SDF files in the
./datafolder. The filtering quality control attribute isGQ, and the filtering condition is(int)value>20. The output is saved to the./folder:java -jar sdfa.jar filter -d ./data -o ./ --filter-gty GQ (int)value>20If the SDF file is a multi-sample file, perform the above-mentioned conditional filtering on the genotypes of each sample in sequence.
To simplify the conversion logic and reduce the storage space of metric field data, we encapsulate common quality control attributes into specific types to facilitate users to operate directly. The details are as follows:
| Storage Class | Storage Type | Value Type | Quality control attributes | Example | Example Explanation |
|---|---|---|---|---|---|
| SingleIntValueBox | int List |
Single int value |
DP,DR,DVGQ,MD,PP |
--filter-gty DP (int)values>10 |
Set the genotype to./ when the DP attribute value is <= 10. |
| TwoIntValueBox | int List |
IntList instance composed of 2 int values |
AD,RA |
--filter-gty AD ((IntList)values).get(0)>10 |
Set the genotype to./ when the first int value of the AD attribute value is <= 10. |
| ThreeIntValueBox | int List |
IntList instance composed of 3 int values |
PL |
--filter-gty PL ((IntList)values).get(0)>10 |
Set the genotype to./ when the first int value of the PL attribute value is <= 10. |
| SingleStringValueBox | string List |
Single string value |
FT others |
--filter-gty FT (string)values.equals(\"TRUE\") |
Set the genotype with the FT attribute not being TR |
SV Filter
Similar to the instructions for genotypes, the instructions for SV filtering are as follows:
--filter-sv <sv_attr> <function>
The above command line filters the specified sv_attr, and function is a JAVA expression that returns a boolean value. Now, let's take an example:
[!NOTE|label:Example 2]
We filter for
IMPRECISEin the INFO field. When this field exists, we filter the SV:java -jar sdfa.jar -d ./data -o ./ --filter-sv IMPRECISE value!=null
In the above example, the sv_attr is IMPRECISE. In the SDFA design, there is a certain order for the positioning and acquisition of sv_attr. The following is the comprehensive filtering order:
| INDEX | Match Field | Value Type | Value Description |
|---|---|---|---|
| 1 | CHR |
string |
Obtain the chromosome name of SV |
| 2 | ID |
Bytes |
Obtain the ID value of SV |
| 3 | REF |
Bytes |
Obtain the REF value of SV |
| 4 | ALT |
Bytes |
Obtain the ALT value of SV |
| 5 | QUAL |
Bytes |
Obtain the QUAL value of SV |
| 6 | FILTER |
Bytes |
Obtain the FILTER value of SV |
| 7 | INFO |
List<Bytes> |
Obtain all INFO field values of SV |
| 8 | FORMAT |
List<Bytes> |
Obtain the quality control values of all genotypes of SV |
| 9 | GT |
CacheGenotypes |
Obtain the genotypes of all samples under SV |
| 10 | LEN |
int |
Obtain the LEN value of SV |
| 11 | SDF_FIELD_NAME |
Object |
Obtain the value of SV in SDF_FIELD_NAME (for details, please refer to the SDF structure) |
| 12 | INFO_ATTR |
Bytes |
Obtain the value of SV in INFO where the KEY is INFO_ATTR |
| 13 | FORMAT_ATTR |
Bytes |
Obtain the value of SV in FORMAT where the KEY is FORMAT_ATTR |
| 14 | UNKNOWN | ERROR | . |
For detailed value retrieval, you can refer to the getV function of SDFRecordWrapper.
[!TIP|label:Filter Order]
It is worth noting that in the actual code execution, the SV Level filtering will be carried out first, and then the Genotype Level filtering.