1. Call SNPs & Infer Phylogeny (CSI Phylogeny)
Rolf Sommer Kaas

2. Single Nucleotide Polymorphism
Assumption: Random + Independent
Find differences (SNP calling), compared to a reference sequence.
Close reference is better
Make pseudo-alignment (independent assumption)
Infer phylogeny

3. CSI Phylogeny – SNP calling
Raw reads
Map reads to reference (BWA)
Call all possible SNPs (Samtools)
Filter positions and SNPs using: coverage, quality, and z-score
Putting ignored snps in “ignored” file.
Z-score = ()/()
Pruning SNPs.
Determine coverage along reference >coverage_file.gz
Output: VCF file (filtered/flt)
Assembled sequence
NUCMER (Part of MUMMER)
Filter based on length from end + pruning
Output: VCF file (filtered/flt)

4. Variant Calling Format (VCF)
##fileformat=VCFv4.1
##..
#CHROM POS ID REF ALT QUAL FILTER
NODE_ C T 73 .
NODE_ G A 125 .
NODE_ A C 222 .
INFO FORMAT qbot_dir/tmp//0807T15624_R1.trim.sorted.bam
DP=1644;VDB=0.0408;AF1=0.5;AC1=1;DP4=2,2,3,14;MQ=60;FQ=63.7;PV4=0.23,4.1e-09,1,0.23 GT:PL:GQ 0/1:103,0,91:94
DP=82;VDB=0.0418;AF1=0.5163;AC1=1;DP4=1,2,19,28;MQ=60;FQ=-15.1;PV4=1,1e-11,1,1 GT:PL:GQ 0/1:155,0,12:15
DP=86;VDB=0.0421;AF1=1;AC1=2;DP4=0,0,40,39;MQ=60;FQ=-265 GT:PL:GQ 1/1:255,238,0:99

5. CSI Phylogeny – Alignment + Tree
Raw reads
Assembled sequence
VCF file (filtered)
Find shared positions in all input
VCF files (final output)
Known accuracy issue
SNP matrix (pairwise)
Concatenate SNPs +
Infer ML phylogeny (FastTree)
Final phylogeny

6. CSI Phylogeny – How to run it
Environment
/home/projects/pr_hackinars/data/19may/load_CSI_phylogeny_env.sh

7. CSI Phylogeny – How to run it
Recommendation: start a screen session
Ks_SNP_tree_pipeline.pl --qbot_dir <dir> --out_dir <dir> --reference <fasta> <arg1> ..<argN>
Options
--disable_qbot_call qbot will not be invoked automatically. You will have to do it yourself with the cmd: qbot –s scripts –e error –o ouput in the qbot dir.
--temp_dir <dir> Location of tmp files. Default is tmp folder inside the qbot_dir.
--filter_depth <int> default: 10.
--filter_relative_depth <int> default: 10. Meaning 10 pct of avg.
--filter_min_depth <int> default: 10. Only applicable if using rel dp option.
--filter_prune <int> default: 10. Meaning 10 SNPs.
--filter_map_quality <int> default: 0. Recommended 25.
--filter_snp_quality <int> default: 0. Recommended 30.
--filter_z <double> default: 0. Recommended 1.96.
--queue_filter_mem <str> default: 3900m. Increase only if your filter job gets killed due to memory.

8. CSI Phylogeny – How to run it