diff --git a/doc/CallerComparisonPlot.pdf b/doc/CallerComparisonPlot.pdf
new file mode 100644
index 00000000..a30922a2
Binary files /dev/null and b/doc/CallerComparisonPlot.pdf differ
diff --git a/test/benchmark/caller_comparison/Snakefile b/test/benchmark/caller_comparison/Snakefile
new file mode 100644
index 00000000..bc2ce45d
--- /dev/null
+++ b/test/benchmark/caller_comparison/Snakefile
@@ -0,0 +1,18 @@
+configfile: "Repos/iGenVar/test/benchmark/caller_comparison/config/config.yaml"
+
+include: "workflow/rules/callers.smk"
+include: "workflow/rules/eval.smk"
+include: "workflow/rules/plots.smk"
+
+##### Target rules #####
+
+rule all:
+    input:
+        # SV calling
+        expand("results/caller_comparison/SVIM/variants.vcf"),
+        # SV lengths
+        # expand("results/caller_comparison/SV-plots/SV-length_SVIM_{minscore}.png", minscore=[3, 5, 7]),
+        # Evaluation
+        expand("results/caller_comparison/eval/results.all.png"),
+        # expand("pipeline/eval/results.tools.{vcf}.png", vcf=VCFS),
+        # expand("pipeline/eval/results.coverages.{vcf}.png", vcf=VCFS),
diff --git a/test/benchmark/caller_comparison/config/config.yaml b/test/benchmark/caller_comparison/config/config.yaml
new file mode 100644
index 00000000..cd726088
--- /dev/null
+++ b/test/benchmark/caller_comparison/config/config.yaml
@@ -0,0 +1,17 @@
+long_bam: data/long_reads/HG002.Sequel.10kb.pbmm2.hs37d5.whatshap.haplotag.RTG.10x.trio_sorted.bam
+long_md_bam: data/long_reads/HG002.Sequel.10kb.pbmm2.hs37d5.whatshap.haplotag.RTG.10x.trio_sorted.md.bam
+long_bai: data/long_reads/HG002.Sequel.10kb.pbmm2.hs37d5.whatshap.haplotag.RTG.10x.trio_sorted.bam.bai
+
+reference_fa_gz: data/reference/hs37d5.fa.gz
+reference_fa: data/reference/hs37d5.fa
+
+parameters:
+    sample: HG002
+    min_var_length: 40
+    max_var_length: 1000000
+
+quality_ranges:
+    igenvar:  {from: 1, to: 80, step: 2}
+    svim:     {from: 1, to: 80, step: 2}
+    sniffles: {from: 1, to: 80, step: 2}
+    pbsv:     {from: 1, to: 80, step: 2}
diff --git a/test/benchmark/caller_comparison/workflow/rules/callers.smk b/test/benchmark/caller_comparison/workflow/rules/callers.smk
new file mode 100644
index 00000000..14866156
--- /dev/null
+++ b/test/benchmark/caller_comparison/workflow/rules/callers.smk
@@ -0,0 +1,161 @@
+wildcard_constraints:
+    sample = config["parameters"]["sample"],
+    min_var_length = config["parameters"]["min_var_length"],
+    max_var_length = config["parameters"]["max_var_length"]
+
+rule run_igenvar:
+    input:
+        bam = config["long_bam"]
+    output:
+        vcf = "results/caller_comparison/iGenVar/variants.vcf"
+    params:
+        min_qual = config["parameters"]["min_qual"]
+    shell:
+        """
+        ./build/iGenVar/bin/iGenVar -t 1 -j {input.bam} -o {output.vcf} \
+        --vcf_sample_name {sample} \
+        --method cigar_string \
+        --method split_read \
+        --min_var_length {min_var_length} \
+        --max_var_length {max_var_length} \
+        --min_qual 2
+        """
+        # Defaults:
+        # --clustering_methods hierarchical_clustering --refinement_methods no_refinement
+        # --max_tol_inserted_length 5 --max_overlap 10 --hierarchical_clustering_cutoff 10
+
+# SVIM
+rule run_svim:
+    input:
+        bam = config["long_bam"],
+        bai = config["long_bai"],
+        genome = config["reference_fa_gz"]
+    output:
+        "results/caller_comparison/SVIM/variants.vcf"
+    resources:
+        mem_mb = 20000,
+        time_min = 600,
+        io_gb = 100
+    params:
+        working_dir = "results/caller_comparison/SVIM/",
+    threads: 1
+    conda:
+        "../../../envs/svim.yaml"
+    shell:
+        """
+        svim alignment --sample {sample} \
+        --partition_max_distance 1000 \
+        --cluster_max_distance 0.5 \
+        --min_sv_size {min_var_length} \
+        --segment_gap_tolerance 20 \
+        --segment_overlap_tolerance 20 \
+        --interspersed_duplications_as_insertions \
+        --tandem_duplications_as_insertions \
+        --read_names \
+        --max_sv_size {max_var_length} \
+        --verbose \
+        {params.working_dir} {input.bam} {input.genome}
+        """
+        # Defaults:
+        # --position_distance_normalizer 900 --edit_distance_normalizer 1.0
+
+# SNIFFLES (we have to loop over min_support, because sniffles does not write a quality score into the vcf)
+rule run_sniffles:
+    input:
+        bam = config["long_md_bam"],
+    output:
+        expand("results/caller_comparison/Sniffles/raw_variants.{minsupport}.vcf",
+               minsupport=list(range(config["quality_ranges"]["sniffles"]["from"],
+                                     config["quality_ranges"]["sniffles"]["to"]+1,
+                                     config["quality_ranges"]["sniffles"]["step"])))
+    resources:
+        mem_mb = 400000,
+        time_min = 1200,
+        io_gb = 100
+    params:
+        qual_from = config["quality_ranges"]["sniffles"]["from"],
+        qual_to = config["quality_ranges"]["sniffles"]["to"]+1,
+        qual_step = config["quality_ranges"]["sniffles"]["step"]
+    threads: 10
+    conda:
+        "../../../envs/sniffles.yaml"
+    shell:
+        """
+        for i in $(seq {params.qual_from} {params.qual_step} {params.qual_to})
+        do
+            sniffles --mapped_reads {input.bam} --vcf results/caller_comparison/Sniffles/raw_variants.$i.vcf \
+            --min_support $i --min_length {min_var_length} --threads {threads} --genotype
+        done
+        """
+
+#see https://github.com/spiralgenetics/truvari/issues/43
+rule fix_sniffles:
+    input:
+        "results/caller_comparison/Sniffles/raw_variants.{support}.vcf"
+    output:
+        "results/caller_comparison/Sniffles/variants.unsorted.min_qual_{support,[0-9]+}.vcf"
+    shell:
+        "sed 's/##INFO=<ID=SUPTYPE,Number=A/##INFO=<ID=SUPTYPE,Number=./' {input} > {output}"
+
+# Split to SV classes
+# Since iGenVar can only find INS and DEL so far, we filter these out for better comparability.
+rule fix_sniffles_2_and_filter_insertions_and_deletions:
+    input:
+        "results/caller_comparison/Sniffles/variants.unsorted.min_qual_{support,[0-9]+}.vcf"
+    output:
+        "results/caller_comparison/Sniffles/variants.min_qual_{support,[0-9]+}.vcf"
+    shell:
+        "bcftools view -i 'SVTYPE=\"DEL\" | SVTYPE=\"INS\"' {input} | bcftools sort > {output}"
+
+#PBSV
+rule run_pbsv_dicsover:
+    input:
+        bam = config["long_bam"]
+    output:
+        svsig_gz = dynamic("results/caller_comparison/pbsv/signatures.{region}.svsig.gz")
+    resources:
+        mem_mb = 400000,
+        time_min = 2000,
+        io_gb = 100
+    threads: 1
+    conda:
+        "../../../envs/pbsv.yaml"
+    shell:
+        # "pbsv discover {input.bam} {output.svsig_gz}"
+        """
+        for i in $(samtools view -H {input.bam} | grep '^@SQ' | cut -f2 | cut -d':' -f2); do
+            pbsv discover --region $i {input.bam} results/caller_comparison/pbsv/signatures.$i.svsig.gz
+        done
+        """
+
+rule run_pbsv_call:
+    input:
+        genome = config["reference_fa"],
+        svsig_gz = dynamic("results/caller_comparison/pbsv/signatures.{region}.svsig.gz")
+    output:
+        vcf = expand("results/caller_comparison/pbsv/variants.min_qual_{minsupport}.vcf",
+                     minsupport=list(range(config["quality_ranges"]["pbsv"]["from"],
+                                           config["quality_ranges"]["pbsv"]["to"]+1,
+                                           config["quality_ranges"]["pbsv"]["step"])))
+    resources:
+        mem_mb = 400000,
+        time_min = 2000,
+        io_gb = 100
+    params:
+        qual_from = config["quality_ranges"]["pbsv"]["from"],
+        qual_to = config["quality_ranges"]["pbsv"]["to"]+1,
+        qual_step = config["quality_ranges"]["pbsv"]["step"]
+    threads: 1
+    conda:
+        "../../../envs/pbsv.yaml"
+    shell:
+        # pbsv call --types DEL,INS,DUP --min-sv-length {params.min_sv_length} --max-ins-length 100K \
+        """
+        for i in $(seq {params.qual_from} {params.qual_step} {params.qual_to})
+        do
+            pbsv call --types DEL,INS --min-sv-length {min_var_length} --max-ins-length 100K \
+            --call-min-reads-all-samples $i --call-min-reads-one-sample $i \
+            --call-min-reads-per-strand-all-samples 0 --call-min-bnd-reads-all-samples 0 --call-min-read-perc-one-sample 0 \
+            --num-threads {threads} {input.genome} {input.svsig_gz} results/caller_comparison/pbsv/variants.min_qual_$i.vcf
+        done
+        """
diff --git a/test/benchmark/caller_comparison/workflow/rules/eval.smk b/test/benchmark/caller_comparison/workflow/rules/eval.smk
new file mode 100644
index 00000000..6f921fe3
--- /dev/null
+++ b/test/benchmark/caller_comparison/workflow/rules/eval.smk
@@ -0,0 +1,83 @@
+rule filter_vcf:
+    input:
+        "results/caller_comparison/{caller,iGenVar|SVIM}/variants.vcf"
+    output:
+        "results/caller_comparison/{caller,iGenVar|SVIM}/variants.min_qual_{min_qual}.vcf"
+    shell:
+        "bcftools view -i 'QUAL>={wildcards.min_qual}' {input} > {output}"
+
+rule bgzip:
+    input:
+        "{name}.vcf"
+    output:
+        "{name}.vcf.gz"
+    shell:
+        "bgzip -c {input} > {output}"
+
+rule tabix:
+    input:
+        "{name}.vcf.gz"
+    output:
+        "{name}.vcf.gz.tbi"
+    shell:
+        "tabix -p vcf {input}"
+
+rule truvari:
+    input:
+        vcf = "results/caller_comparison/{caller}/variants.min_qual_{min_qual}.vcf.gz",
+        index = "results/caller_comparison/{caller}/variants.min_qual_{min_qual}.vcf.gz.tbi"
+    params:
+        output_dir = "results/caller_comparison/eval/{caller}/min_qual_{min_qual}"
+    output:
+        summary = "results/caller_comparison/eval/{caller}/min_qual_{min_qual}/summary.txt"
+    conda:
+        "../../../envs/truvari.yaml"
+    shell:
+        """
+        rm -rf {params.output_dir} && truvari bench -b data/truth_set/HG002_SVs_Tier1_v0.6.vcf.gz \
+        -c {input.vcf} -o {params.output_dir} --passonly --includebed data/truth_set/HG002_SVs_Tier1_v0.6.bed -p 0
+        """
+
+rule reformat_truvari_results:
+    input:
+        "results/caller_comparison/eval/{caller}/min_qual_{min_qual}/summary.txt"
+    output:
+        "results/caller_comparison/eval/{caller}/min_qual_{min_qual}/pr_rec.txt"
+    threads: 1
+    shell:
+        """
+        cat {input} | grep '\<precision\>\|\<recall\>' | tr -d ',' |sed 's/^[ \t]*//' | tr -d '\"' | tr -d ' ' \
+        | tr ':' '\t' | awk 'OFS=\"\\t\" {{ print \"{wildcards.caller}\", \"{wildcards.min_qual}\", $1, $2 }}' > {output}
+        """
+
+rule cat_truvari_results_all:
+    input:
+        igenvar  = expand("results/caller_comparison/eval/iGenVar/min_qual_{min_qual}/pr_rec.txt",
+                          min_qual=list(range(config["quality_ranges"]["igenvar"]["from"],
+                                              config["quality_ranges"]["igenvar"]["to"]+1,
+                                              config["quality_ranges"]["igenvar"]["step"]))),
+        svim     = expand("results/caller_comparison/eval/SVIM/min_qual_{min_qual}/pr_rec.txt",
+                          min_qual=list(range(config["quality_ranges"]["svim"]["from"],
+                                              config["quality_ranges"]["svim"]["to"]+1,
+                                              config["quality_ranges"]["svim"]["step"]))),
+        sniffles = expand("results/caller_comparison/eval/Sniffles/min_qual_{min_qual}/pr_rec.txt",
+                          min_qual=list(range(config["quality_ranges"]["sniffles"]["from"],
+                                              config["quality_ranges"]["sniffles"]["to"]+1,
+                                              config["quality_ranges"]["sniffles"]["step"]))),
+        pbsv = expand("results/caller_comparison/eval/pbsv/min_qual_{min_qual}/pr_rec.txt",
+                          min_qual=list(range(config["quality_ranges"]["pbsv"]["from"],
+                                              config["quality_ranges"]["pbsv"]["to"]+1,
+                                              config["quality_ranges"]["pbsv"]["step"])))
+    output:
+        igenvar  = temp("results/caller_comparison/eval/igenvar.all_results.txt"),
+        svim     = temp("results/caller_comparison/eval/svim.all_results.txt"),
+        sniffles = temp("results/caller_comparison/eval/sniffles.all_results.txt"),
+        pbsv     = temp("results/caller_comparison/eval/pbsv.all_results.txt"),
+        all = "results/caller_comparison/eval/all_results.txt"
+    threads: 1
+    run:
+        shell("cat {input.igenvar} > {output.igenvar}")
+        shell("cat {input.svim} > {output.svim}")
+        shell("cat {input.sniffles} > {output.sniffles}")
+        shell("cat {input.pbsv} > {output.pbsv}")
+        shell("cat {output.igenvar} {output.svim} {output.sniffles} {output.pbsv} > {output.all}")
diff --git a/test/benchmark/caller_comparison/workflow/rules/plots.smk b/test/benchmark/caller_comparison/workflow/rules/plots.smk
new file mode 100644
index 00000000..2f951bb6
--- /dev/null
+++ b/test/benchmark/caller_comparison/workflow/rules/plots.smk
@@ -0,0 +1,12 @@
+rule plot_pr_all_results:
+    input:
+        "results/caller_comparison/eval/all_results.txt"
+    output:
+        "results/caller_comparison/eval/results.all.png"
+    log:
+        "logs/rplot.all.log"
+    shell:
+        """
+        Rscript --vanilla Repos/iGenVar/test/benchmark/caller_comparison/workflow/scripts/plot_all_results.R \
+        {input} {output} > {log}
+        """
diff --git a/test/benchmark/caller_comparison/workflow/scripts/plot_all_results.R b/test/benchmark/caller_comparison/workflow/scripts/plot_all_results.R
new file mode 100644
index 00000000..d88fb701
--- /dev/null
+++ b/test/benchmark/caller_comparison/workflow/scripts/plot_all_results.R
@@ -0,0 +1,26 @@
+library(tidyverse)
+library(scales)
+
+args = commandArgs(trailingOnly=TRUE)
+
+res <- read_tsv(args[1], col_names = c("caller", "min_qual", "metric", "value"))
+res$caller = factor(res$caller,
+                    levels=c('iGenVar', 'SVIM', 'Sniffles', 'pbsv'),
+                    labels=c('iGenVar', 'SVIM', 'Sniffles', 'pbsv'))
+res %>%
+    filter(metric %in% c("recall", "precision")) %>%
+    pivot_wider(names_from=metric, values_from=value) %>%
+    filter(recall!=0 | precision!=0) %>%
+    mutate(precision = 100*precision, recall = 100*recall) %>%
+    ggplot(aes(recall, precision, color=caller, pch=caller)) +
+      geom_point(size=0.5) +
+      # scale_shape_manual(values=c(15,16,17)) +
+      # scale_color_manual(values=c("deepskyblue3", "goldenrod2", "firebrick2")) +
+      geom_path() +
+      labs(y = "Precision", x = "Recall", color = "Tool", pch = "Tool") +
+      lims(x=c(0,100), y=c(0,100)) +
+      theme_bw() +
+      theme(panel.spacing = unit(0.75, "lines")) +
+      theme(text = element_text(size=14), axis.text.x = element_text(size=9), axis.text.y = element_text(size=9))
+
+ggsave(args[2], width=20, height=12)
diff --git a/test/benchmark/envs/cyvcf2.yaml b/test/benchmark/envs/cyvcf2.yaml
new file mode 100644
index 00000000..ed48896a
--- /dev/null
+++ b/test/benchmark/envs/cyvcf2.yaml
@@ -0,0 +1,8 @@
+channels:
+ - bioconda
+dependencies:
+ - python>=3.6
+ - pip
+ - pip:
+   - cyvcf2
+   - matplotlib
diff --git a/test/benchmark/shell_scripts/environment.yml b/test/benchmark/envs/environment.yml
similarity index 100%
rename from test/benchmark/shell_scripts/environment.yml
rename to test/benchmark/envs/environment.yml
diff --git a/test/benchmark/envs/pbsv.yaml b/test/benchmark/envs/pbsv.yaml
new file mode 100644
index 00000000..028d22e5
--- /dev/null
+++ b/test/benchmark/envs/pbsv.yaml
@@ -0,0 +1,4 @@
+channels:
+ - bioconda
+dependencies:
+ - pbsv=2.6.2
diff --git a/test/benchmark/envs/samtools.yaml b/test/benchmark/envs/samtools.yaml
new file mode 100644
index 00000000..1e13e14f
--- /dev/null
+++ b/test/benchmark/envs/samtools.yaml
@@ -0,0 +1,5 @@
+channels:
+ - bioconda
+#Sorting BAM files from NGMLR fails in samtools version>=1.10, therefore fix to version 1.9
+dependencies:
+ - samtools=1.9
diff --git a/test/benchmark/envs/sniffles.yaml b/test/benchmark/envs/sniffles.yaml
new file mode 100644
index 00000000..01e2fdce
--- /dev/null
+++ b/test/benchmark/envs/sniffles.yaml
@@ -0,0 +1,4 @@
+channels:
+ - bioconda
+dependencies:
+ - sniffles=1.0.11
diff --git a/test/benchmark/envs/svim.yaml b/test/benchmark/envs/svim.yaml
new file mode 100644
index 00000000..0d692987
--- /dev/null
+++ b/test/benchmark/envs/svim.yaml
@@ -0,0 +1,4 @@
+channels:
+ - bioconda
+dependencies:
+ - svim=1.4.2
diff --git a/test/benchmark/envs/truvari.yaml b/test/benchmark/envs/truvari.yaml
new file mode 100644
index 00000000..68bcaed8
--- /dev/null
+++ b/test/benchmark/envs/truvari.yaml
@@ -0,0 +1,7 @@
+channels:
+ - bioconda
+dependencies:
+ - python>=3.6
+ - pip
+ - pip:
+   - Truvari
diff --git a/test/benchmark/envs/truvari_environment.yaml b/test/benchmark/envs/truvari_environment.yaml
deleted file mode 100644
index 807f9d14..00000000
--- a/test/benchmark/envs/truvari_environment.yaml
+++ /dev/null
@@ -1,8 +0,0 @@
-channels:
-  - defaults
-  - conda-forge
-  - bioconda
-dependencies:
-  - python=3
-  - pip:
-    - truvari
diff --git a/test/benchmark/shell_scripts/iGenVar_init.sh b/test/benchmark/iGenVar_init.sh
similarity index 62%
rename from test/benchmark/shell_scripts/iGenVar_init.sh
rename to test/benchmark/iGenVar_init.sh
index 9caff0ad..f3d28928 100755
--- a/test/benchmark/shell_scripts/iGenVar_init.sh
+++ b/test/benchmark/iGenVar_init.sh
@@ -20,7 +20,7 @@ git clone https://github.com/seqan/iGenVar.git
 cd iGenVar
 git submodule update --recursive --init
 
-echo "$(tput setaf 1)$(tput setab 7)------- iGenVar downloaded (1/5) --------$(tput sgr 0)" 1>&3
+echo "$(tput setaf 1)$(tput setab 7)------- iGenVar downloaded (1/7) --------$(tput sgr 0)" 1>&3
 
 cd ../..
 mkdir -p build && cd build && mkdir -p iGenVar && cd iGenVar
@@ -30,7 +30,7 @@ make -j 16
 
 make test && make doc
 
-echo "$(tput setaf 1)$(tput setab 7)------- iGenVar built (2/5) --------$(tput sgr 0)" 1>&3
+echo "$(tput setaf 1)$(tput setab 7)------- iGenVar built (2/7) --------$(tput sgr 0)" 1>&3
 
 # -------- -------- get data and unzip -------- -------- #
 cd ../..
@@ -40,14 +40,31 @@ mkdir -p data && cd data
 # wget ...
 # cd ..
 
-echo "$(tput setaf 1)$(tput setab 7)------- short reads downloaded (3/5) --------$(tput sgr 0)" 1>&3
+echo "$(tput setaf 1)$(tput setab 7)------- short reads downloaded (3/7) --------$(tput sgr 0)" 1>&3
 
 mkdir -p long_reads && cd long_reads
 wget --retry-connrefused --waitretry=30 --read-timeout=30 --timeout=30 --tries=20 --no-clobber --no-verbose \
     https://ftp-trace.ncbi.nlm.nih.gov/ReferenceSamples/giab/data/AshkenazimTrio/HG002_NA24385_son/PacBio_CCS_10kb/alignment/HG002.Sequel.10kb.pbmm2.hs37d5.whatshap.haplotag.RTG.10x.trio.bam
 
-echo "$(tput setaf 1)$(tput setab 7)------- long reads downloaded (4/5) --------$(tput sgr 0)" 1>&3
+wget --retry-connrefused --waitretry=30 --read-timeout=30 --timeout=30 --tries=20 --no-clobber --no-verbose \
+    https://ftp-trace.ncbi.nlm.nih.gov/ReferenceSamples/giab/data/AshkenazimTrio/HG002_NA24385_son/PacBio_CCS_10kb/alignment/HG002.Sequel.10kb.pbmm2.hs37d5.whatshap.haplotag.RTG.10x.trio.bam.bai
+cd ..
+echo "$(tput setaf 1)$(tput setab 7)------- long reads downloaded (4/7) --------$(tput sgr 0)" 1>&3
+
+# -------- -------- get reference ../../data -------- -------- #
+mkdir -p reference && cd reference
+
+wget --retry-connrefused --waitretry=30 --read-timeout=30 --timeout=30 --tries=20 --no-clobber --no-verbose \
+    ftp://ftp-trace.ncbi.nih.gov/1000genomes/ftp/technical/reference/phase2_reference_assembly_sequence/hs37d5.fa.gz
+gzip --decompress --keep hs37d5.fa.gz
+cd ..
+echo "$(tput setaf 1)$(tput setab 7)------- reference downloaded (5/7) --------$(tput sgr 0)" 1>&3
+
+samtools calmd -b data/long_reads/HG002.Sequel.10kb.pbmm2.hs37d5.whatshap.haplotag.RTG.10x.trio_sorted.bam \
+> data/long_reads/HG002.Sequel.10kb.pbmm2.hs37d5.whatshap.haplotag.RTG.10x.trio_sorted.md.bam \
+data/reference/GCA_000001405.28_GRCh38.p13_genomic.fna.gz
 
+echo "$(tput setaf 1)$(tput setab 7)------- missing MD tags added (6/7) --------$(tput sgr 0)" 1>&3
 
 # -------- -------- get truth set ../../data -------- -------- #
 mkdir -p truth_set && cd truth_set
@@ -58,12 +75,18 @@ wget --retry-connrefused --waitretry=30 --read-timeout=30 --timeout=30 --tries=2
     https://ftp-trace.ncbi.nlm.nih.gov/ReferenceSamples/giab/../../data/AshkenazimTrio/analysis/NIST_SVs_Integration_v0.6/HG002_SVs_Tier1_v0.6.vcf.gz.tbi
 wget --retry-connrefused --waitretry=30 --read-timeout=30 --timeout=30 --tries=20 --no-clobber --no-verbose \
     https://ftp-trace.ncbi.nlm.nih.gov/ReferenceSamples/giab/../../data/AshkenazimTrio/analysis/NIST_SVs_Integration_v0.6/HG002_SVs_Tier1_v0.6.bed
+cd ../..
+echo "$(tput setaf 1)$(tput setab 7)------- truth set downloaded (7/7) --------$(tput sgr 0)" 1>&3
 
-echo "$(tput setaf 1)$(tput setab 7)------- truth set downloaded (5/5) --------$(tput sgr 0)" 1>&3
+mkdir -p results
 
+# -------- -------- pre installation steps -------- -------- #
+# We do our benchmarks with snakemake, and need several tools like
+# python3, samtools, snakemake, tabix, pip and truvari
+# we recommend to use miniconda: https://docs.conda.io/en/latest/miniconda.html
+# and run
 
-cd ../..
-mkdir -p results
+conda env create -f Repos/iGenVar/test/benchmark/envs/environment.yml
 
 # ----------------------------------------
 echo "$(tput setaf 1)$(tput setab 7)------- Initial steps - done --------$(tput sgr 0)" 1>&3
diff --git a/test/benchmark/Snakefile b/test/benchmark/parameter_benchmarks/Snakefile
similarity index 96%
rename from test/benchmark/Snakefile
rename to test/benchmark/parameter_benchmarks/Snakefile
index 9fdf45fa..23a693e0 100644
--- a/test/benchmark/Snakefile
+++ b/test/benchmark/parameter_benchmarks/Snakefile
@@ -55,7 +55,8 @@ rule truvari:
         output_dir = "results/truvari/{parameter_name}/{parameter_value}_min_qual_{min_qual}"
     output:
         summary = "results/truvari/{parameter_name}/{parameter_value}_min_qual_{min_qual}/summary.txt"
-    # conda: "envs/truvari_environment.yaml"
+    conda:
+        "../envs/truvari.yaml"
     shell:
         """
         rm -rf {params.output_dir} && truvari bench -b data/truth_set/HG002_SVs_Tier1_v0.6.vcf.gz \
@@ -110,4 +111,4 @@ rule plot_pr_all_results:
     output:
         "results/plots/{parameter_name}.results.all.png"
     shell:
-        "Rscript --vanilla Repos/iGenVar/test/benchmark/scripts/plot_all.R {input} {wildcards.parameter_name} {output}"
+        "Rscript --vanilla Repos/iGenVar/test/benchmark/parameter_benchmarks/plot_all.R {input} {wildcards.parameter_name} {output}"
diff --git a/test/benchmark/shell_scripts/iGenVar_run_benchmark.sh b/test/benchmark/parameter_benchmarks/iGenVar_run_benchmark.sh
similarity index 95%
rename from test/benchmark/shell_scripts/iGenVar_run_benchmark.sh
rename to test/benchmark/parameter_benchmarks/iGenVar_run_benchmark.sh
index ac0a094a..eead516c 100755
--- a/test/benchmark/shell_scripts/iGenVar_run_benchmark.sh
+++ b/test/benchmark/parameter_benchmarks/iGenVar_run_benchmark.sh
@@ -38,7 +38,7 @@ else
    echo "Use one core."
 fi
 
-/usr/bin/time -v snakemake --snakefile Repos/iGenVar/test/benchmark/Snakefile --cores $cores \ # --quiet --forcerun run_igenvar \
+/usr/bin/time -v snakemake --snakefile Repos/iGenVar/test/benchmark/parameter_benchmarks/Snakefile --cores $cores \ # --quiet --forcerun run_igenvar \
 --rerun-incomplete --stats logs/${current_date}_snakemake_stats.txt --timestamp --use-conda
 
 echo "$(tput setaf 1)$(tput setab 7)------- done: iGenVar with cigar string & split read method --------$(tput sgr 0)" 1>&3
diff --git a/test/benchmark/scripts/plot_all.R b/test/benchmark/parameter_benchmarks/plot_all.R
similarity index 100%
rename from test/benchmark/scripts/plot_all.R
rename to test/benchmark/parameter_benchmarks/plot_all.R