diff --git a/CHANGELOG.md b/CHANGELOG.md
index 4b0cac1..a5b6d1c 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -4,6 +4,15 @@ This file contains all notable changes to Bambu-Pipe.
 
 ---
 
+## [v0.10.3] - 2026-10-01
+
+### Changed
+- Increased the minimum adapter overlap in the `PREPROCESS_FASTQ` adapter re-search from the cutadapt default (3 bp) to 10 bp, reducing the number of valid reads discarded due to short chance matches at read ends
+
+### Fixed
+- Corrected the 10x3v3 and 10x3v4 end adapter (`rev_primer_f`/`rev_primer_r` in `adapter_seq_config.csv`) to `CCCATGTACTCTGCGTTGATACCACTGCTT`, matching 10x3v2 and Visium; the previous sequence does not occur in these reads, so the adapter was rarely trimmed
+- Resolved an edge case where `PREPROCESS_FASTQ` failed if the input FASTQ filename had the same length as the barcode (16) or UMI (10/12) pattern; the unquoted `?` wildcards in the flexiplex pattern were expanded by the shell to match the filename, which was not intended
+
 ## [v0.10.2] - 2026-09-23
 
 ### Fixed
diff --git a/assets/10x_config/adapter_seq_config.csv b/assets/10x_config/adapter_seq_config.csv
index 839572c..5fd3525 100644
--- a/assets/10x_config/adapter_seq_config.csv
+++ b/assets/10x_config/adapter_seq_config.csv
@@ -1,7 +1,7 @@
 technology,fwd_primer_f,fwd_primer_r,rev_primer_f,rev_primer_r,TSO_f,TSO_r
 10x3v2,CTACACGACGCTCTTCCGATCT,AGATCGGAAGAGCGTCGTGTAG,CCCATGTACTCTGCGTTGATACCACTGCTT,AAGCAGTGGTATCAACGCAGAGTACATGGG,,
-10x3v3,CTACACGACGCTCTTCCGATCT,AGATCGGAAGAGCGTCGTGTAG,CCCCTCTGCGTTGATACCACTGCTT,AAGCAGTGGTATCAACGCAGAGGGG,,
-10x3v4,CTACACGACGCTCTTCCGATCT,AGATCGGAAGAGCGTCGTGTAG,CCCCTCTGCGTTGATACCACTGCTT,AAGCAGTGGTATCAACGCAGAGGGG,,
+10x3v3,CTACACGACGCTCTTCCGATCT,AGATCGGAAGAGCGTCGTGTAG,CCCATGTACTCTGCGTTGATACCACTGCTT,AAGCAGTGGTATCAACGCAGAGTACATGGG,,
+10x3v4,CTACACGACGCTCTTCCGATCT,AGATCGGAAGAGCGTCGTGTAG,CCCATGTACTCTGCGTTGATACCACTGCTT,AAGCAGTGGTATCAACGCAGAGTACATGGG,,
 10x5v2,CTACACGACGCTCTTCCGATCT,AGATCGGAAGAGCGTCGTGTAG,GTACTCTGCGTTGATACCACTGCTT,AAGCAGTGGTATCAACGCAGAGTAC,TTTCTTATATGGG,CCCATATAAGAAA
 10x5v3,CTACACGACGCTCTTCCGATCT,AGATCGGAAGAGCGTCGTGTAG,GTACTCTGCGTTGATACCACTGCTT,AAGCAGTGGTATCAACGCAGAGTAC,TTTCTTATATGGG,CCCATATAAGAAA
 visium-v1,CTACACGACGCTCTTCCGATCT,AGATCGGAAGAGCGTCGTGTAG,CCCATGTACTCTGCGTTGATACCACTGCTT,AAGCAGTGGTATCAACGCAGAGTACATGGG,,
diff --git a/modules/preprocess_fastq.nf b/modules/preprocess_fastq.nf
index 9bf4d27..2943718 100644
--- a/modules/preprocess_fastq.nf
+++ b/modules/preprocess_fastq.nf
@@ -64,7 +64,7 @@ process PREPROCESS_FASTQ {
             flexiplex_f=$params.flexiplex_f_3prime
         fi
 
-        flexiplex -p $task.cpus -k \$barcode_list \$flank_seq -f \$flexiplex_f -e $params.flexiplex_e
+        flexiplex -p $task.cpus -k \$barcode_list "\${flank_seq[@]}" -f \$flexiplex_f -e $params.flexiplex_e
     }
 
     # Trims reverse primer with cutadapt
@@ -72,7 +72,8 @@ process PREPROCESS_FASTQ {
         cutadapt -a \$rev_primer_f --cores $task.cpus -
     }
 
-    # Re-searches all primer and TSO sequences to remove non-standard reads 
+    # Re-searches all primer and TSO sequences to remove non-standard reads
+    # (--overlap 10: require >=10 bp of adapter overlap so we do not discard too many reads)
     cutadapt_re_search() {
         if [[ $meta.chemistry == 10x5* ]]; then
             args="-b \$fwd_primer_f -b \$fwd_primer_r -b \$TSO_f -b \$TSO_r -b \$rev_primer_f -b \$rev_primer_r"
@@ -80,7 +81,7 @@ process PREPROCESS_FASTQ {
             args="-b \$fwd_primer_f -b \$fwd_primer_r -b \$rev_primer_f -b \$rev_primer_r"
         fi
         
-        cutadapt \$args --action none --discard --cores $task.cpus -
+        cutadapt \$args --overlap 10 --action none --discard --cores $task.cpus -
     }
 
     # Reverse complements FASTQ reads to orient them in the transcript direction (for 3' and visium chemistries only)
@@ -102,7 +103,8 @@ process PREPROCESS_FASTQ {
     # Extract chemistry specfic parameters from config files
     IFS=',' read -r _ left_flank barcode umi right_flank < <(awk -F',' -v chem=$meta.chemistry '\$1 == chem' $flank_seq_config)
     IFS=',' read -r _ fwd_primer_f fwd_primer_r rev_primer_f rev_primer_r TSO_f TSO_r < <(awk -F',' -v chem=$meta.chemistry '\$1 == chem' $adapter_seq_config)
-    flank_seq="-x \$left_flank -b \$barcode -u \$umi -x \$right_flank"
+    # Store flexiplex search pattern as an array so the '?' wildcards are not expanded as shell globs
+    flank_seq=(-x "\$left_flank" -b "\$barcode" -u "\$umi" -x "\$right_flank")
 
     # Flexiplex filter (barcode filtering)
     # For visium chemistries, search the whole rank curve instead
@@ -112,7 +114,7 @@ process PREPROCESS_FASTQ {
     else
         max_rank_arg=""
     fi
-    flexiplex -p $task.cpus \$flank_seq -f 0 ${sample}_chopper_out.fastq
+    flexiplex -p $task.cpus "\${flank_seq[@]}" -f 0 ${sample}_chopper_out.fastq
     flexiplex-filter \$max_rank_arg -w $whitelist --outfile my_filtered_barcode_list.txt flexiplex_barcodes_counts.txt
 
     # Chain commands: Flexiplex demultiplexing -> (Optional) Save intermediate file after flexiplex -> Cutadapt trimming of reverse primer -> \
diff --git a/nextflow.config b/nextflow.config
index 690c916..22b638f 100644
--- a/nextflow.config
+++ b/nextflow.config
@@ -2,7 +2,7 @@ manifest {
     name        = 'Bambu-Pipe'
     description = 'Context-aware transcript quantification from long-read single-cell and spatial transcriptomics data'
     author      = 'Andre Sim, Chin Hao Lee, Min Hao Ling'
-    version     = 'v0.10.2'
+    version     = 'v0.10.3'
     mainScript  = 'main.nf'
     nextflowVersion = '!>=26.04.0'
 }