Using yaml file instead of named env to create absplice scores

deeprvat/annotations/annotations.py

@@ -1210,9 +1210,10 @@ def merge_abscores(
     annotations = annotations.rename(columns={"Gene": "gene_id"})
     annotations.drop_duplicates(inplace=True, subset=["gene_id", "id"])
     original_len = len(annotations)
-
+    all_ids  = set(annotations.id)
+    all_absplice_scores.drop_duplicates(subset = ["chrom", "pos", "ref", "alt", "gene_id"], inplace = True)
     logger.info("Merging")
-    merged = pd.merge(
+    annotations = pd.merge(
         annotations,
         all_absplice_scores,
         validate="1:1",
@@ -1221,24 +1222,25 @@ def merge_abscores(
     )
 
     logger.info("Sanity checking merge")
-    assert len(merged) == original_len
+    assert len(annotations) == original_len
+    assert set(annotations.id) == all_ids
     logger.info(
-        f"len of merged after dropping duplicates: {len(merged.drop_duplicates(subset=['id', 'gene_id']))}"
+        f"len of merged after dropping duplicates: {len(annotations.drop_duplicates(subset=['id', 'gene_id']))}"
     )
-    logger.info(f"len of merged without dropping duplicates: {len(merged)}")
+    logger.info(f"len of merged without dropping duplicates: {len(annotations)}")
 
-    assert len(merged.drop_duplicates(subset=["id", "gene_id"])) == len(merged)
+    assert len(annotations.drop_duplicates(subset=["id", "gene_id"])) == len(annotations)
 
     logger.info(
-        f'Filling {merged["AbSplice_DNA"].isna().sum()} '
+        f'Filling {annotations["AbSplice_DNA"].isna().sum()} '
         "missing AbSplice values with 0"
     )
-    merged["AbSplice_DNA"] = merged["AbSplice_DNA"].fillna(0)
+    annotations["AbSplice_DNA"] = annotations["AbSplice_DNA"].fillna(0)
 
     annotation_out_file = out_file
 
     logger.info(f"Writing to {annotation_out_file}")
-    merged.to_parquet(annotation_out_file, engine="pyarrow")
+    annotations.to_parquet(annotation_out_file, engine="pyarrow")
 
 
 pd.options.mode.chained_assignment = None

pipelines/annotations.snakefile

@@ -124,7 +124,6 @@ kipoi_repo_dir = Path(config["kipoiveff_repo_dir"])
 ncores_addis = int(config.get("n_jobs_addids") or 32)
 
 # init absplice
-absplice_repo_dir = Path(config["absplice_repo_dir"])
 n_cores_absplice = int(config.get("n_cores_absplice") or 4)
 ncores_merge_absplice = int(config.get("n_cores_merge_absplice") or 8)
 ncores_agg_absplice = int(config.get("ncores_agg_absplice") or 4)
@@ -154,7 +153,7 @@ file_stems = [
 
 absplice_download_dir = (
     config.get("absplice_download_dir")
-    or absplice_repo_dir / "example" / "data" / "resources" / "downloaded_files"
+    or anno_tmp_dir / "absplice" 
 )
 absplice_output_dir = config.get("absplice_output_dir", anno_tmp_dir / "absplice")
 vcf_id = anno_tmp_dir / "{file_stem}"

pipelines/config/deeprvat_annotation_config.yaml

@@ -19,7 +19,6 @@ anno_dir : output_dir/annotations
 
 vep_cache_dir : repo_dir/ensembl-vep/cache/
 vep_plugin_dir : repo_dir/ensembl-vep/Plugins
-absplice_repo_dir : repo_dir/absplice
 deeprvat_repo_dir : ../..
 kipoiveff_repo_dir : repo_dir/kipoi-veff2
 faatpipe_repo_dir : repo_dir/faatpipe

pipelines/resources/absplice.yaml

@@ -29,4 +29,4 @@ dependencies:
     - interpret == 0.2.7
     - interpret-core == 0.2.7
     - git+https://github.com/gagneurlab/splicemap.git@9e9831f32c221e850e26757a5f1c132dcd565640
-    - git+https://github.com/gagneurlab/absplice.git@e4060e63cca074ac938ea28850a38290c7e8b198
+    - -e git+https://github.com/gagneurlab/absplice.git@5d8b3fbe3ab1cc6ebd3a68857a50b48671753362#egg=absplice

pipelines/resources/absplice_download.snakefile

@@ -53,7 +53,7 @@ rule download_human_fasta:
     output:
         Path(absplice_download_dir) / config_download["fasta"][genome]["file"],
     conda:
-        "absplice"
+        "./absplice.yaml"
     shell:
         "wget -O - {params} | gunzip -c > {output}"
 
@@ -75,7 +75,7 @@ rule download_splicemaps:
         splicemap_psi5=Path(absplice_download_dir)
         / config_download["splicemap"]["psi5"],
     conda:
-        "absplice"
+        "./absplice.yaml"
     shell:
         "splicemap_download --version {params.version} --splicemap_dir {params.dirname} --tissues {wildcards.tissue}"
 
@@ -103,7 +103,7 @@ if absplice_main_conf["AbSplice_RNA"] == True:
         output:
             Path(absplice_download_dir) / config_download["gtf"][genome]["file"],
         conda:
-            "absplice"
+            "./absplice.yaml"
         shell:
             "wget -O - {params} | gunzip -c > {output}"
 
@@ -119,7 +119,7 @@ if absplice_main_conf["AbSplice_RNA"] == True:
             coding_genes=Path(absplice_download_dir)
             / config_download["gtf"][genome]["coding_genes"],
         conda:
-            "absplice"
+            "./absplice.yaml"
         resources:
             mem_mb=lambda wildcards, attempt: attempt * 16000,
         script:

pipelines/resources/absplice_splicing_pred_DNA.snakefile

@@ -33,7 +33,7 @@ rule mmsplice_splicemap:
         mem_mb=30_000,
         threads=4,
     conda:
-        "absplice"
+        "./absplice.yaml"
     output:
         result=Path(absplice_output_dir)
         / config_pred["splicing_pred"]["mmsplice_splicemap"],
@@ -113,7 +113,7 @@ else:
             spliceai_csv=Path(absplice_output_dir)
             / config_pred["splicing_pred"]["spliceai"],
         conda:
-            "absplice"
+            "./absplice.yaml"
         run:
             from absplice.utils import read_spliceai_vcf
 
@@ -130,7 +130,7 @@ rule absplice_dna:
     params:
         extra_info=absplice_main_conf["extra_info_dna"],
     conda:
-        "absplice"
+        "./absplice.yaml"
     output:
         absplice_dna=absplice_output_dir
         / "{genome}"

pipelines/resources/environment_spliceai_rocksdb.yaml

@@ -4,19 +4,21 @@ channels:
 dependencies:
   - python==3.9.15
   - python-rocksdb==0.7.0
-  - rocksdb==6.13.3
-  - tensorflow-gpu==2.6.2
-  - cudnn==8.8.0.121
-  - cudatoolkit==11.0.3
-  - pyarrow==6.0.1
-  - tqdm==4.65.0
-  - pip==23.2.1
-  - setuptools==68.0.0
-  - click==8.1.6
-  - pooch==1.7.0
-  - bioconda::snakemake==7.31.0
+  - rocksdb==6.29.5
+  - tensorflow-gpu==2.14.0
+  - cudnn==8.9.7.29
+  - cudatoolkit==11.8.0
+  - pyarrow==12.0.1
+  - tqdm==4.66.4
+  - pip==24.0
+  - pandas=2.2.2
+  - setuptools==70.0.0
+  - click==8.1.7
+  - pooch==1.8.1
+  - numpy=1.26.4
+  - bioconda::snakemake==7.32.4
   - bioconda::kipoiseq==0.7.1
   - bioconda::spliceai==1.3.1
-  - bioconda::cyvcf2==0.30.16
+  - bioconda::cyvcf2==0.30.28
   - pip:
     - git+https://github.com/gagneurlab/spliceai_rocksdb.git@3c40d6e61b19e907e802c979c76d52ea5c41c1d5

tests/annotations/test_annotations.py

@@ -385,6 +385,13 @@ def test_aggregate_abscores(
             "vep_deepripe_deepsea.parquet",
             "vep_deepripe_deepsea_absplice.parquet",
         ),
+        (
+            "merge_absplice_scores_exons",
+            "abSplice_score_file.parquet",
+            "vep_deepripe_deepsea.parquet",
+            "vep_deepripe_deepsea_absplice.parquet",
+        ),
+        
     ],
 )
 def test_merge_absplice_scores(