parameter and methods update

kishwarshafin · Mar 16, 2022 · 3dd53a8 · 3dd53a8
1 parent e77f54f
commit 3dd53a8
Show file tree

Hide file tree

Showing 2 changed files with 37 additions and 16 deletions.
diff --git a/docs/misc/pepper_methods.md b/docs/misc/pepper_methods.md
@@ -71,6 +71,21 @@ iii) `find_candidates`: This step takes the predictions of the previous step to
   </p>
   * All variants are used for haplotagging with margin.
 
+
+Update in `r0.8`: From this version, we split the SNP and INDEL candidates into two VCFs to enable SNP and INDEL calling separately with DeepVariant.
+The PEPPER output VCFs are:
+```bash
+PEPPER_VARIANT_FULL.vcf.gz : Contails all variants reported by PEPPER.
+PEPPER_VARIANT_OUTPUT_PEPPER.vcf.gz: Contains only the variants that are found high-quality in PEPPER and will not be re-gotyped with DeepVariant.
+PEPPER_VARIANT_OUTPUT_VARIANT_CALLING_INDEL.vcf.gz: Contains INDEL candidate variants that will be re-genotyped with DeepVariant.
+PEPPER_VARIANT_OUTPUT_VARIANT_CALLING_SNPs.vcf.gz: Contains SNP candidate variants that will be re-genotyped with DeepVariant.
+PEPPER_VARIANT_OUTPUT_VARIANT_CALLING.vcf.gz: Contains all candidate variants (SNPs + INDELs) that will be re-genotyped with DeepVariant.
+```
+
+For Oxford Nanopore variant calling, we now use:
+* DeepVariant `rows` mode that uses INDEL realignment for calling candidate variants from `PEPPER_VARIANT_OUTPUT_VARIANT_CALLING_INDEL.vcf.gz`.
+* DeepVariant `none` mode that uses SNP calling for calling candidate variants from `PEPPER_VARIANT_OUTPUT_VARIANT_CALLING_SNPs.vcf.gz`.
+
 ### Step 2: Margin [Haplotagging reads]
 Margin uses the read-based evidence of linkage between heterozygous variant sites to find the most likely assignment of reads and alleles to haplotypes.
 The tool first selects a set of high-quality primary reads and variants for use in the initial phasing workflow.
@@ -111,7 +126,7 @@ The following parameters are used to select reads:
 * `phase.maxDepth`: reads are downsampled to an expected depth of this parameter.  Longer reads are preferentially selected during downsampling.
 
 ### Step 3: DeepVariant
-DeepVariant re-genotypes the candidates proposed by `PEPPER-HP` using reads haplotagged by `Margin`. We use the following features for DeepVariant:
+DeepVariant re-genotypes the candidates proposed by `PEPPER` using reads haplotagged by `Margin`. We use the following features for DeepVariant:
 <p align="center">
 <img src="img/deepvariant_pileup.png" alt="DeepVariant summary" height=100px>
 </p>

diff --git a/docs/usage/usage_and_parameters.md b/docs/usage/usage_and_parameters.md
@@ -103,18 +103,24 @@ run_pepper_margin_deepvariant call_variant --help
 | --margin_phase_model       | String | Path to a custom margin model.               |
 
 ### Parameters for DeepVariant
-| Parameter                    | Type    | Description                                                                        |
-|------------------------------|---------|------------------------------------------------------------------------------------|
-| --dv_model                   | String  | Path to a custom DeepVariant model.                                                |
-| --dv_alt_aligned_pileup      | String  | DeepVariant alt_align_pileup used for make_examples. [none, rows, diff_channels]   |
-| --dv_realign_reads           | String  | If true then local read alingment will be performed. [set: true/false]             |
-| --dv_partition_size          | String  | DeepVariant partition_size used for make_examples.                                 |
-| --dv_min_mapping_quality     | Integer | DeepVariant minimum mapping quality.                                               |
-| --dv_min_base_quality        | Integer | DeepVariant minimum base quality.                                                  |
-| --dv_sort_by_haplotypes      | String  | If true then haplotype sorting will be used. [set: true/false]                     |
-| --dv_parse_sam_aux_fields    | String  | If true then auxiliary field parsing is enabled. [set: true/false]                 |
-| --dv_add_hp_channel          | String  | If true then hp channel will be added. [set: true/false]                           |
-| --dv_use_hp_information      | String  | If true then hp information will be properly used. [set: true/false]               |
-| --dv_use_multiallelic_mode   | String  | If true multiallelic model will be used during post-processing. [set: true/false]  |
-| --dv_vsc_min_fraction_indels | String  | DeepVariant minimum vsc fraction for indels.                                       |
-| --dv_vsc_min_fraction_snps   | String  | DeepVariant minimum vsc fraction for snps.                                         |
+| Parameter                     | Type    | Description                                                                                              |
+|-------------------------------|---------|----------------------------------------------------------------------------------------------------------|
+| --dv_model                    | String  | Path to a custom DeepVariant model. If set, then this model will be used for both SNP and INDEL calling. |
+| --dv_model_snp                | String  | Custom DeepVariant model for SNP calling.                                                                |
+| --dv_model_indel              | String  | Custom DeepVariant model for INDEL calling.                                                              |
+| --dv_alt_aligned_pileup       | String  | alt_align_pileup used for make_examples of --dv_model parameter. [none, rows, diff_channels]             |
+| --dv_alt_aligned_pileup_snp   | String  | alt_align_pileup used for make_examples of --dv_model_snp parameter. [none, rows, diff_channels]         |
+| --dv_alt_aligned_pileup_indel | String  | alt_align_pileup used for make_examples of --dv_model_indel parameter. [none, rows, diff_channels]       |
+| --dv_pileup_image_width       | String  | DeepVariant image width. [none, rows, diff_channels]                                                     |
+| --dv_realign_reads            | String  | If true then local read alingment will be performed. [set: true/false]                                   |
+| --dv_partition_size           | String  | DeepVariant partition_size used for make_examples.                                                       |
+| --dv_min_mapping_quality      | Integer | DeepVariant minimum mapping quality.                                                                     |
+| --dv_min_base_quality         | Integer | DeepVariant minimum base quality.                                                                        |
+| --dv_vsc_min_fraction_indels  | String  | DeepVariant minimum vsc fraction for indels.                                                             |
+| --dv_vsc_min_fraction_snps    | String  | DeepVariant minimum vsc fraction for snps.                                                               |
+| --dv_sort_by_haplotypes       | String  | If true then haplotype sorting will be used. [set: true/false]                                           |
+| --dv_parse_sam_aux_fields     | String  | If true then auxiliary field parsing is enabled. [set: true/false]                                       |
+| --dv_add_hp_channel           | String  | If true then hp channel will be added. [set: true/false]                                                 |
+| --dv_use_hp_information       | String  | If true then hp information will be properly used. [set: true/false]                                     |
+| --dv_use_multiallelic_mode    | String  | If true multiallelic model will be used during post-processing. [set: true/false]                        |
+