diff --git a/workflow/envs/paired_end_read_merger.yaml b/workflow/envs/paired_end_read_merger.yaml
index 2bccdb22..8b655505 100644
--- a/workflow/envs/paired_end_read_merger.yaml
+++ b/workflow/envs/paired_end_read_merger.yaml
@@ -16,6 +16,7 @@ dependencies:
   - scipy
   - matplotlib-base
   - mafft
+  - git-lfs
   - pip
   - pip:
     - git+https://github.com/cbg-ethz/smallgenomeutilities@dev
diff --git a/workflow/rules/snv.smk b/workflow/rules/snv.smk
index 5fd4bc73..462e337f 100644
--- a/workflow/rules/snv.smk
+++ b/workflow/rules/snv.smk
@@ -295,14 +295,21 @@ rule paired_end_read_merger:
             if config.viloca["consensus"]
             else reference_file
         ),
+        fname_ref_idx=(
+            cohortdir("cohort_consensus.fasta")
+            if config.viloca["consensus"]
+            else reference_file
+        )+".fai",
     output:
-        fname_sam_merged=f"results/{{sample}}/alignment/REF_aln.merged.sam",
+        fname_bam_merged="{dataset}/alignment/REF_aln.merged.bam",
+        fname_sam_merged=temp_with_prefix("{dataset}/alignment/REF_aln.merged.sam"),
+        fname_sam=temp_with_prefix("{dataset}/alignment/REF_aln.sam"),
+        fname_sam_nonmerged="{dataset}/alignment/REF_aln.nonmerged.sam",
+        fname_sam_sort=temp_with_prefix("{dataset}/alignment/REF_aln.sort.sam"),
     params:
-        fname_sam=temp(f"results/{{sample}}/alignment/REF_aln.sam"),
-        fname_sam_nonmerged=f"results/{{sample}}/alignment/REF_aln.nonmerged.sam",
-        fname_sam_sort=temp(f"results/{{sample}}/alignment/REF_aln.sort.sam"),
         SAMTOOLS=config.applications["samtools"],
         PAIRED_END_READ_MERGER=config.applications["paired_end_read_merger"],
+        sort_tmp=temp_prefix("{dataset}.tmp"),
     log:
         outfile="{dataset}/alignment/paired_end_read_merger.out.log",
         errfile="{dataset}/alignment/paired_end_read_merger.err.log",
@@ -311,12 +318,17 @@ rule paired_end_read_merger:
     shell:
         """
         ## Preparation
-        fname_reference_idx = "${input.fname_reference_idx}.fai"
-        {params.SAMTOOLS} view -h -T {input.fname_reference} -t ${fname_reference_idx} {input.fname_bam} > {params.fname_sam}
+        {params.SAMTOOLS} view -h -T {input.fname_ref} -t {input.fname_ref_idx} {input.fname_bam} -o {output.fname_sam} > {log.outfile} 2> >(tee {log.errfile} >&2)
         ## sort accrording to QNAME
-        {params.SAMTOOLS} sort -T tmp -O sam -n {params.fname_sam} > {params.fname_sam_sort}
+        rm -f '{params.sort_tmp}'.[0-9]*.bam
+        {params.SAMTOOLS} sort -T "{params.sort_tmp}" -O sam -n {output.fname_sam} -o {output.fname_sam_sort} >> {log.outfile} 2> >(tee -a {log.errfile} >&2)
         ## run script
-        {params.PAIRED_END_READ_MERGER} {input.fname_sam_sort} {output.fname_sam_merged} {params.fname_sam_nonmerged} {input.fname_ref}
+        {params.PAIRED_END_READ_MERGER} {output.fname_sam_sort} {output.fname_sam_merged} {output.fname_sam_nonmerged} {input.fname_ref} >> {log.outfile} 2> >(tee -a {log.errfile} >&2)
+        touch {output.fname_sam_nonmerged}
+        ## sort
+        rm -f '{params.sort_tmp}'.[0-9]*.bam
+        {params.SAMTOOLS} sort -T "{params.sort_tmp}" -o "{output.fname_bam_merged}" "{output.fname_sam_merged}" >> {log.outfile} 2> >(tee -a {log.errfile} >&2)
+        {params.SAMTOOLS} index "{output.fname_bam_merged}" >> {log.outfile} 2> >(tee -a {log.errfile} >&2)
         """
 
 
@@ -328,7 +340,7 @@ rule viloca:
             else reference_file
         ),
         BAM=(
-            rules.paired_end_read_merger.output.fname_sam_merged,
+            rules.paired_end_read_merger.output.fname_bam_merged
             if config.viloca["merge_paired_end_reads"]
             else alignment_wildcard
         ),
diff --git a/workflow/schemas/config_schema.json b/workflow/schemas/config_schema.json
index cc25d855..797c9a3b 100644
--- a/workflow/schemas/config_schema.json
+++ b/workflow/schemas/config_schema.json
@@ -1237,6 +1237,12 @@
                     "description": "Indicate whether to use the cohort-consensus sequence from the analyzed samples (output from `minor_variants` rule located in the cohort-wide output `results/cohort_onsensus.fasta`) or the reference sequence by setting this option to False.",
                     "examples": [false]
                 },
+                "merge_paired_end_reads": {
+                    "type": "boolean",
+                    "default": false,
+                    "description": "Merge paired-end reads in the preprocessing. This is a preprocessing snakemake rule.",
+                    "examples": [false]
+                },
                 "shift": {
                     "type": "integer",
                     "default": 3,