475424b5d722b4ff761681f1d3bd91ad666fb2d6
angie
  Wed Apr 1 10:17:03 2020 -0700
For nested clades, add all clade samples to per-clade VCF.  Also generate Newick (.nh) tree file and tweak sample name separator, in preparation for drawing the nextstrain tree in the VCF left label area.  refs #25188

diff --git src/hg/utils/otto/nextstrainNcov/doUpdate.sh src/hg/utils/otto/nextstrainNcov/doUpdate.sh
index cd735e1..299f077 100755
--- src/hg/utils/otto/nextstrainNcov/doUpdate.sh
+++ src/hg/utils/otto/nextstrainNcov/doUpdate.sh
@@ -1,65 +1,67 @@
 #!/bin/bash
 set -beEu -o pipefail
 
 #	Do not modify this script, modify the source tree copy:
 #	kent/src/hg/utils/nextstrainNcov/doUpdate.sh
 
 ottoDir=/hive/data/outside/otto/nextstrainNcov
 chromSizes=/hive/data/genomes/wuhCor1/chrom.sizes
 gbdbDir=/gbdb/wuhCor1/nextstrain
 
 cd $ottoDir
 
 # The file is named .json, but is actually gzipped, so gunzip it.
 curl -s http://data.nextstrain.org/ncov.json | gunzip -c > ncov.json
 curl -s -I http://data.nextstrain.org/ncov.json \
 | grep Last-Mod | sed -re 's/Last-Modified: //; s/\r//;' \
    > ncov.json.date
 latestDate=$(cat ncov.json.date)
 oldDate=$(cat old.ncov.json.date)
 if [ $(date -d "$latestDate" +%s) -le $(date -d "$oldDate" +%s)  ]; then
     exit 0
 fi
 
 today=`date +%F`
 mkdir -p $today
 cp -p ncov.json $today
 mv ncov.json old.ncov.json
 mv ncov.json.date old.ncov.json.date
 
 runDir=$ottoDir/$today
 cd $runDir
 
 #Generate bed and VCF files
 $ottoDir/nextstrain.py
 
 # bgzip & tabix the VCF files
 bgzip -f nextstrainSamples.vcf
 tabix -p vcf nextstrainSamples.vcf.gz
 for clade in A1a A2 A2a A3 A6 A7 B B1 B2 B4; do
   bgzip -f nextstrainSamples$clade.vcf
   tabix -p vcf nextstrainSamples$clade.vcf.gz
 done
 
 # bigBed-ify the gene names and "clades"
 bedToBigBed -type=bed4 -tab -verbose=0 nextstrainGene.bed $chromSizes \
     nextstrainGene.bb
 
 sort -k2n,2n nextstrainClade.bed > nextstrainClade.sorted.bed
 bedToBigBed -as=$ottoDir/nextstrainClade.as -type=bed12+7 -tab -verbose=0 \
     nextstrainClade.sorted.bed $chromSizes \
     nextstrainClade.bb
 
 # Archive
 mkdir -p $ottoDir/archive/$today
 cp -p $runDir/nextstrainGene.bb $runDir/nextstrainClade.bb \
     $runDir/nextstrainSamples*.vcf.gz{,.tbi} \
+    $runDir/nextstrain.nh \
+    $runDir/ncov.json \
     $ottoDir/archive/$today
-cp -p $runDir/ncov.json $ottoDir/archive/$today
 
 # Install
 ln -sf $runDir/nextstrainGene.bb $runDir/nextstrainClade.bb \
     $runDir/nextstrainSamples*.vcf.gz{,.tbi} \
+    $runDir/nextstrain.nh \
     $gbdbDir/
 
 echo "Updated nextstrain/ncov `date` (ncov.json date $latestDate)"