merge branch develop

Author: suhrig

README.md

@@ -88,7 +88,6 @@ Please refer to the [user manual](http://arriba.readthedocs.io/en/latest/) for i
    - [RefSeq annotation](https://arriba.readthedocs.io/en/latest/current-limitations/#refseq-annotation)
    - [Memory consumption](https://arriba.readthedocs.io/en/latest/current-limitations/#memory-consumption)
    - [Adapter trimming](https://arriba.readthedocs.io/en/latest/current-limitations/#adapter-trimming)
-   - [Small insert size](https://arriba.readthedocs.io/en/latest/current-limitations/#small-insert-size)
    - [Viral detection](https://arriba.readthedocs.io/en/latest/current-limitations/#viral-detection)
    - [Targeted sequencing](https://arriba.readthedocs.io/en/latest/current-limitations/#targeted-sequencing)
 

download_references.sh

@@ -1,24 +1,24 @@
 #!/bin/bash
 
 declare -A ASSEMBLIES
-ASSEMBLIES[hs37d5]="ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/technical/reference/phase2_reference_assembly_sequence/hs37d5.fa.gz"
+ASSEMBLIES[hs37d5]="http://ftp.1000genomes.ebi.ac.uk/vol1/ftp/technical/reference/phase2_reference_assembly_sequence/hs37d5.fa.gz"
 ASSEMBLIES[hg19]="http://hgdownload.cse.ucsc.edu/goldenpath/hg19/bigZips/chromFa.tar.gz"
-ASSEMBLIES[GRCh37]="ftp://ftp.ensembl.org/pub/grch37/release-87/fasta/homo_sapiens/dna/Homo_sapiens.GRCh37.dna.primary_assembly.fa.gz"
+ASSEMBLIES[GRCh37]="http://ftp.ensembl.org/pub/grch37/release-87/fasta/homo_sapiens/dna/Homo_sapiens.GRCh37.dna.primary_assembly.fa.gz"
 ASSEMBLIES[hg38]="http://hgdownload.cse.ucsc.edu/goldenpath/hg38/bigZips/hg38.chromFa.tar.gz"
-ASSEMBLIES[GRCh38]="ftp://ftp.sanger.ac.uk/pub/cancer/dockstore/human/GRCh38_hla_decoy_ebv/core_ref_GRCh38_hla_decoy_ebv.tar.gz"
+ASSEMBLIES[GRCh38]="http://ftp.ensembl.org/pub/release-93/fasta/homo_sapiens/dna/Homo_sapiens.GRCh38.dna.primary_assembly.fa.gz"
 ASSEMBLIES[mm10]="http://hgdownload.cse.ucsc.edu/goldenpath/mm10/bigZips/chromFa.tar.gz"
-ASSEMBLIES[GRCm38]="ftp://ftp.ensembl.org/pub/release-99/fasta/mus_musculus/dna/Mus_musculus.GRCm38.dna.primary_assembly.fa.gz"
+ASSEMBLIES[GRCm38]="http://ftp.ensembl.org/pub/release-99/fasta/mus_musculus/dna/Mus_musculus.GRCm38.dna.primary_assembly.fa.gz"
 ASSEMBLIES[mm39]="http://hgdownload.cse.ucsc.edu/goldenpath/mm39/bigZips/mm39.chromFa.tar.gz"
 ASSEMBLIES[GRCm39]="http://ftp.ensembl.org/pub/release-104/fasta/mus_musculus/dna/Mus_musculus.GRCm39.dna.primary_assembly.fa.gz"
 
 declare -A ANNOTATIONS
-ANNOTATIONS[GENCODE19]="ftp://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_human/release_19/gencode.v19.annotation.gtf.gz"
+ANNOTATIONS[GENCODE19]="http://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_human/release_19/gencode.v19.annotation.gtf.gz"
 ANNOTATIONS[RefSeq_hg19]="http://hgdownload.cse.ucsc.edu/goldenpath/hg19/database/refGene.txt.gz"
-ANNOTATIONS[ENSEMBL87]="ftp://ftp.ensembl.org/pub/grch37/release-87/gtf/homo_sapiens/Homo_sapiens.GRCh37.87.chr.gtf.gz"
-ANNOTATIONS[GENCODE28]="ftp://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_human/release_28/gencode.v28.annotation.gtf.gz"
+ANNOTATIONS[ENSEMBL87]="http://ftp.ensembl.org/pub/grch37/release-87/gtf/homo_sapiens/Homo_sapiens.GRCh37.87.chr.gtf.gz"
+ANNOTATIONS[GENCODE28]="http://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_human/release_28/gencode.v28.annotation.gtf.gz"
 ANNOTATIONS[RefSeq_hg38]="http://hgdownload.cse.ucsc.edu/goldenpath/hg38/database/refGene.txt.gz"
-ANNOTATIONS[ENSEMBL93]="ftp://ftp.ensembl.org/pub/release-93/gtf/homo_sapiens/Homo_sapiens.GRCh38.93.chr.gtf.gz"
-ANNOTATIONS[GENCODEM25]="ftp://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_mouse/release_M25/gencode.vM25.annotation.gtf.gz"
+ANNOTATIONS[ENSEMBL93]="http://ftp.ensembl.org/pub/release-93/gtf/homo_sapiens/Homo_sapiens.GRCh38.93.chr.gtf.gz"
+ANNOTATIONS[GENCODEM25]="http://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_mouse/release_M25/gencode.vM25.annotation.gtf.gz"
 ANNOTATIONS[RefSeq_mm10]="http://hgdownload.cse.ucsc.edu/goldenpath/mm10/database/refGene.txt.gz"
 ANNOTATIONS[GENCODEM26]="http://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_mouse/release_M26/gencode.vM26.annotation.gtf.gz"
 ANNOTATIONS[RefSeq_mm39]="http://hgdownload.cse.ucsc.edu/goldenpath/mm39/database/refGene.txt.gz"
@@ -82,7 +82,7 @@ else
 fi |
 if [ "$VIRAL" = "viral" ]; then
 	# drop viral contigs from assembly
-	awk '/^>/{ contig=$1 } contig!~/^>NC_|^>AC_|^>chrEBV/{ print }'
+	awk '/^>/{ contig=$1 } contig!~/^>NC_|^>AC_/{ print }'
 else
 	cat
 fi > "$ASSEMBLY$VIRAL.fa"