diff --git a/Dockerfile b/Dockerfile
index 085692b..e58df99 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -19,8 +19,6 @@ COPY pyproject.toml ./
 COPY etc ./etc
 COPY sbin ./sbin
 COPY src ./src
-COPY loading/data/splign-manual ./splign-manual
-COPY loading ./loading
 RUN pip install -e .[dev]
 
 
diff --git a/README.md b/README.md
index 2ade8c5..a05be31 100644
--- a/README.md
+++ b/README.md
@@ -357,6 +357,19 @@ docker compose run seqrepo-load
 UTA_ETL_SKIP_GENE_LOAD=true docker compose run uta-load
 ```
 
+#### 2C. Manual splign transcripts
+To load splign-manual transcripts, the workflow expects an input txdata.yaml file and splign alignments. Define this path 
+using the environment variable $UTA_SPLIGN_MANUAL_DIR:
+- `$UTA_SPLIGN_MANUAL_DIR/splign-manual/txdata.yaml`
+- `$UTA_SPLIGN_MANUAL_DIR/splign-manual/alignments/*.splign`
+
+[txdata.yaml](loading/data/splign-manual/txdata.yaml) defines the transcripts and their metadata. The [alignments dir](loading/data/splign-manual/alignments) contains the splign alignments.
+To run the workflow:
+```
+export UTA_SPLIGN_MANUAL_DIR=$(pwd)/loading/data/splign-manual/
+docker compose run splign-manual
+```
+
 UTA has updated and the database has been dumped into a pgd file in `UTA_ETL_WORK_DIR`. SeqRepo has been updated in place.
 
 
diff --git a/docker-compose.yml b/docker-compose.yml
index 671227b..295f961 100644
--- a/docker-compose.yml
+++ b/docker-compose.yml
@@ -62,13 +62,14 @@ services:
     network_mode: host
   splign-manual:
     image: uta-update
-    command: sbin/uta-splign-manual ${UTA_ETL_OLD_UTA_VERSION} /uta-splign-manual/work /uta-splign-manual/logs
+    command: sbin/uta-splign-manual ${UTA_ETL_OLD_UTA_VERSION} /uta-splign-manual/input /uta-splign-manual/work /uta-splign-manual/logs
     depends_on:
       uta:
         condition: service_healthy
     volumes:
       - ${UTA_ETL_NCBI_DIR}:/ncbi-dir
       - ${UTA_ETL_SEQREPO_DIR}:/usr/local/share/seqrepo
+      - ${UTA_SPLIGN_MANUAL_DIR}:/uta-splign-manual/input
       - ${UTA_ETL_WORK_DIR}:/uta-splign-manual/work
       - ${UTA_ETL_LOG_DIR}:/uta-splign-manual/logs
     network_mode: host
diff --git a/loading/data/splign-manual/README.md b/loading/data/splign-manual/README.md
index 3a4d62b..f364281 100644
--- a/loading/data/splign-manual/README.md
+++ b/loading/data/splign-manual/README.md
@@ -50,7 +50,7 @@ For a given RefSeq transcript (e.g., NM_000996.3), do the following:
    - Click on the gene id to go to the gene page (e.g., `6165`)
    - N.B. Strand is inferred from the orientation of aligned exons.
 
-1. Enter the gene and CDS info in txdata.yaml
+1. Enter the gene, geneID, and CDS info in txdata.yaml
 
 1. Get the chromosome and coordinates from the gene page
    - From the "Genomic Context" section, note the chromosomal
diff --git a/loading/data/splign-manual/generate-loading-data b/sbin/generate-loading-data
similarity index 100%
rename from loading/data/splign-manual/generate-loading-data
rename to sbin/generate-loading-data
diff --git a/sbin/uta-splign-manual b/sbin/uta-splign-manual
index 06a0062..4357b3f 100755
--- a/sbin/uta-splign-manual
+++ b/sbin/uta-splign-manual
@@ -5,12 +5,13 @@
 set -euxo pipefail
 
 source_uta_v=$1
-working_dir=$2
-log_dir=$3
+input_dir=$2
+working_dir=$3
+log_dir=$4
 
-if [ -z "$working_dir" ] || [ -z "$log_dir" ]
+if [ -z "$input_dir" ] || [ -z "$working_dir" ] || [ -z "$log_dir" ]
 then
-    echo 'Usage: sbin/uta-splign-manual <source_uta_version> <working_dir> <log_dir>'
+    echo 'Usage: sbin/uta-splign-manual <source_uta_version> <input_dir> <working_dir> <log_dir>'
     exit 1
 fi
 
@@ -22,7 +23,7 @@ mkdir -p "$log_dir"
 mkdir -p "$working_dir"
 
 # Generate txinfo.gz and exonset.gz files
-python loading/data/splign-manual/generate-loading-data loading/data/splign-manual/alignments/*.splign 2>&1 --txdata loading/data/splign-manual/txdata.yaml --output-dir $working_dir | tee "$log_dir/generate-loading-data.log"
+python sbin/generate-loading-data $input_dir/alignments/*.splign 2>&1 --txdata $input_dir/txdata.yaml --output-dir $working_dir | tee "$log_dir/generate-loading-data.log"
 
 # Generate fasta files
 seqrepo export $(gzip -cdq $working_dir/txinfo.gz  | cut -f2 | tail +2) | gzip -c > $working_dir/seqs.fa.gz | tee "$log_dir/seqrepo-export.log"