Hello, which exome BED file for intervals do you recommend for human GRCh37 genome? I'm using the GATK one, that has 189894 intervals (SNP_GRCh37_hg19_woChr.bed), less than Ensembl, NCBI and UCSC ones. But when sarek starts the BaseRecalibrator process it increases to 811010. I think this is why the pipeline slows downs A LOT, both in mapping and recalibration. Thanks in advance! { "genome": "GATK.GRCh37", "input": "samplesheet.csv", "outdir": "/storage/scratch01/groups/ce/BLCA_Germline/Results/Wang", "intervals": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Annotation/intervals/Broad.human.exome.b37.interval_list", "wes": true, "save_output_as_bam": true, "igenomes_ignore": true, "igenomes_base":"", "snpeff_cache":"", "vep_cache":"", "ascat_genome": "hg19", "ascat_alleles": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Annotation/ASCAT/G1000_alleles_hg19.zip", "ascat_loci": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Annotation/ASCAT/G1000_loci_hg19.zip", "ascat_loci_gc": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Annotation/ASCAT/GC_G1000_hg19.zip", "ascat_loci_rt": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Annotation/ASCAT/RT_G1000_hg19.zip", "bwa": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Sequence/BWAIndex/", "chr_dir": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Sequence/Chromosomes", "dbsnp": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Annotation/GATKBundle/dbsnp_138.b37.vcf.gz", "dbsnp_tbi": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Annotation/GATKBundle/dbsnp_138.b37.vcf.gz.tbi", "dict": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Sequence/WholeGenomeFasta/human_g1k_v37_decoy.dict", "fasta": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Sequence/WholeGenomeFasta/human_g1k_v37_decoy.fasta", "fasta_fai": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Sequence/WholeGenomeFasta/human_g1k_v37_decoy.fasta.fai", "germline_resource": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Annotation/GATKBundle/af-only-gnomad.raw.sites.vcf.gz", "germline_resource_tbi": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Annotation/GATKBundle/af-only-gnomad.raw.sites.vcf.gz.tbi", "known_indels": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Annotation/GATKBundle/{1000G_phase1,Mills_and_1000G_gold_standard}.indels.b37.vcf.gz", "known_indels_tbi": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Annotation/GATKBundle/{1000G_phase1,Mills_and_1000G_gold_standard}.indels.b37.vcf.gz.tbi", "known_snps": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Annotation/GATKBundle/1000G_phase1.snps.high_confidence.b37.vcf.gz", "known_snps_tbi": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Annotation/GATKBundle/1000G_phase1.snps.high_confidence.b37.vcf.gz.tbi", "mappability": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Annotation/Control-FREEC/out100m2_hg19.gem", "msisensor2_models": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Annotation/MSIsensor2/models_hg19_GRCh37/", "msisensorpro_scan": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Annotation/MSIsensorPro/human_g1k_v37_decoy.msisensor_scan.list", "ngscheckmate_bed": "/storage/scratch01/groups/ce/BLCA_Germline/GRCh37/references/Homo_sapiens/GATK/GRCh37/Annotation/NGSCheckMate/SNP_GRCh37_hg19_woChr.bed" }

Whole exome sequencing relies on an enrichment of exon sequences (typically by hybridization), which is done with a kit. Each kit has certain target sequences, and kits are different from each other (certainly kits from different manufacturers, but also different kits [or versions of a kit] from the same manufacturer). There is no "standard" intervals file -- you need to use the one that shows the exon sequences targeted by the specific kit that was used for your samples. You may need to ask whoever made the libraries which kit was used (maybe even get the catalog number to be sure) and then get the correct BED file from the manufacturers website (make sure you not get the correct one for the kit, but also for the genome assembly [GRCh37 or 38] you want to use; also check the chromosome naming convention [

chr1

vs

1

, etc.]). (I have only worked with Agilent kits so far, but I might be able to help find the correct one for that. If in doubt, you can e-mail the manufactures with kit description and ideally the catalog number and ask which BED file(s) to use.) Often there are several BED files that are slightly different, e.g. for many Agilent kits, there are files for • the "designed" targets, i.e. the exons they wanted to target, and which were used as "baits", i.e. the sequences that were actually used in the hybridization (Agilent calls that "regions"), • the areas that are actually covered by those baits (some might hybridize with reasonable stringency to more than one locus in the genome, so this is typically larger than the designed targets; Agilent calls that "covered"), • and because reads that are captured by hybridization may only partially overlap the targets, there is typically a "padded" file, where the covered intervals are extended on each side by a number of bases (e.g. 50 or 100). (There may also be other files with combined or additional information.) Note that many (Agilent) BED files contain headers with UCSC "track" information at the top, which need to be removed to avoid errors in sarek. I normally use the "padded" file when I run sarek, but I might use the design targets ("regions") to filter the output, especially for tumor mutational burden calculations. As mentioned in your previous thread, the insane proliferation of processes can be caused by a misinterpretation of extra columns, so it is safest to remove them (this goes for any BED file with more than 3 columns).

Thanks so much Thomas for your extensive explanation, which I found very clarifying. I have 3 columns in the BED file. I will check regarding the manufacturer BED file corresponding to the library kit used.