diff --git a/aws-params.yaml b/aws-params.yaml
index ba0ae3e..5c2aeea 100644
--- a/aws-params.yaml
+++ b/aws-params.yaml
@@ -1,8 +1,9 @@
-readsdir: "s3://vkc-nextflow/rawfastq/"
+readsdir: "s3://vkc-nextflow/scratch/"
 outdir: "s3://vkc-nextflow/output/"
 human_genome: "s3://biobakery-databases/kneaddata_databases/"
-metaphlan_db: "s3://biobakery-databases/metaphlan_databases/"
+metaphlan_db: "s3://biobakery-databases/metaphlan_v4_databases/"
 humann_bowtie_db: "s3://biobakery-databases/humann_databases/chocophlan"
 humann_protein_db: "s3://biobakery-databases/humann_databases/uniref"
 humann_utility_db: "s3://biobakery-databases/humann_databases/utility_mapping"
-filepattern: "*_L00{1,2,3,4}_R{1,2}_001.fastq.gz"
\ No newline at end of file
+filepattern: "*_{1,2}.fastq.gz"
+sra_list: "s3://vkc-nextflow/sra_accessions.txt"
\ No newline at end of file
diff --git a/main.nf b/main.nf
index e99896d..32fdf79 100755
--- a/main.nf
+++ b/main.nf
@@ -14,14 +14,8 @@ workflow {
 
     human_genome      = params.human_genome
     metaphlan_db      = params.metaphlan_db
-    humann_bowtie_db  = params.humann_bowtie_db
-    humann_protein_db = params.humann_protein_db
-    humann_utility_db = params.humann_utility_db
     
     knead_out     = kneaddata(read_pairs_ch, human_genome)
-    metaphlan_out = metaphlan(knead_out[0], knead_out[1], metaphlan_db)
+    metaphlan_out = metaphlan(knead_out[0], metaphlan_db)
     metaphlan_bzip = metaphlan_bzip(metaphlan_out[0], metaphlan_out[4])
-    humann_out    = humann(metaphlan_out[0], metaphlan_out[1], metaphlan_out[2], humann_bowtie_db, humann_protein_db)
-    regroup_out   = humann_regroup(humann_out[0], humann_out[1], humann_utility_db)
-    humann_rename(regroup_out, humann_utility_db)
 }
diff --git a/metaphlan_only.nf b/metaphlan_only.nf
new file mode 100644
index 0000000..e4f02fc
--- /dev/null
+++ b/metaphlan_only.nf
@@ -0,0 +1,18 @@
+#!/usr/bin/env nextflow
+nextflow.enable.dsl=2
+
+include { metaphlan; metaphlan_bzip } from './processes/metaphlan.nf'
+
+workflow {
+    
+    read_pairs_ch = Channel
+        .fromFilePairs(
+            [ "$params.readsdir/$params.filepattern",
+              "$params.readsdir/*_kneaddata.fastq.gz" ],
+            size:-1)
+
+    metaphlan_db      = params.metaphlan_db
+
+    metaphlan_out = metaphlan(read_pairs_ch, metaphlan_db)
+    metaphlan_bzip = metaphlan_bzip(metaphlan_out[0], metaphlan_out[4])
+}
diff --git a/nextflow.config b/nextflow.config
index 0097e86..25c1d71 100644
--- a/nextflow.config
+++ b/nextflow.config
@@ -46,6 +46,15 @@ profiles {
         process {
 
             executor = 'awsbatch'
+
+            withName: prefetch_and_split {
+                maxForks = 1
+                memory = '8.G'
+                time   = '2.h'
+                cpus  =  2
+                container = 'public.ecr.aws/j5i5h1i5/bzip2sra:mamba-v1.0'
+                queue = 'Nextflow-IOPS'                
+            }
     
             withName: kneaddata {
                 maxForks = 4
@@ -59,10 +68,10 @@ profiles {
 
             withName: metaphlan {
                 maxForks = 4
-                memory = '8.G'
+                memory = '32.G'
                 time   = '12.h'
                 cpus  =  8
-                container = 'public.ecr.aws/j5i5h1i5/metaphlan-nodb:mamba-v3.1'
+                container = 'public.ecr.aws/j5i5h1i5/metaphlan-nodb:mamba-v4.1'
                 queue = 'Nextflow-metaphlan'                
             }
 
diff --git a/processes/metaphlan.nf b/processes/metaphlan.nf
index e081e44..47ba0b7 100644
--- a/processes/metaphlan.nf
+++ b/processes/metaphlan.nf
@@ -3,9 +3,8 @@ process metaphlan {
     publishDir "$params.outdir/metaphlan", pattern: "{*.tsv}"
 
     input:
-    tuple val(sample), path(kneads)
-    path unmatched
-    path metaphlan_db
+        tuple val(sample), path(kneads)
+        path metaphlan_db
 
     output:
     val  sample                  , emit: sample
@@ -15,13 +14,9 @@ process metaphlan {
     path "${sample}.sam"
 
     script:
-    def forward = kneads[0]
-    def reverse = kneads[1]
-    def unf = unmatched[0]
-    def unr = unmatched[1]
-
+    
     """
-    cat $forward $reverse $unf $unr > ${sample}_grouped.fastq.gz
+    cat $kneads > ${sample}_grouped.fastq.gz
     
     metaphlan ${sample}_grouped.fastq.gz ${sample}_profile.tsv \
         --bowtie2out ${sample}_bowtie2.tsv \
@@ -49,4 +44,4 @@ process metaphlan {
     """
     bzip2 -v $sam
     """
-}
\ No newline at end of file
+}
diff --git a/processes/prefetch_and_split.nf b/processes/prefetch_and_split.nf
new file mode 100644
index 0000000..6a4bd06
--- /dev/null
+++ b/processes/prefetch_and_split.nf
@@ -0,0 +1,24 @@
+process prefetch_and_split {
+    tag "$sra_id"
+
+    input:
+    val sra_id
+
+    output:
+    tuple val(sra_id), path("${sra_id}_*.fastq.gz")
+
+    publishDir "${params.readsdir}", mode: 'copy'
+
+    script:
+    """
+    echo "Prefetching $sra_id to S3"
+    prefetch $sra_id -v -v
+
+    echo "Splitting $sra_id into FASTQ files directly on S3"
+    fasterq-dump $sra_id --split-files --threads ${task.cpus} -v -v
+
+    echo "Compressing FASTQ files"
+    gzip ${sra_id}_1.fastq
+    gzip ${sra_id}_2.fastq
+    """
+}
\ No newline at end of file
diff --git a/tubular.nf b/tubular.nf
new file mode 100644
index 0000000..631af1a
--- /dev/null
+++ b/tubular.nf
@@ -0,0 +1,26 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl=2
+
+include { prefetch_and_split } from './processes/prefetch_and_split.nf'
+include { kneaddata } from './processes/kneaddata.nf'
+include { metaphlan; metaphlan_bzip } from './processes/metaphlan.nf'
+include { humann; humann_regroup; humann_rename } from './processes/humann.nf'
+
+workflow {
+    
+    // Step 1: Read the SRA accession list and create a channel
+    sra_ch = Channel.fromPath(params.sra_list)
+        .splitText()
+        .map { it.trim() }
+
+    // Step 2: Prefetch and split FASTQ files
+    fastq_pairs_ch = prefetch_and_split(sra_ch)
+    
+    human_genome      = params.human_genome
+    metaphlan_db      = params.metaphlan_db
+    
+    knead_out     = kneaddata(fastq_pairs_ch, human_genome)
+    metaphlan_out = metaphlan(knead_out[0], metaphlan_db)
+    metaphlan_bzip = metaphlan_bzip(metaphlan_out[0], metaphlan_out[4])
+}