updated configs

FlorianBarkmann · FlorianBarkmann · commit c1a098239da4 · 2025-08-11T10:46:08.000+02:00
diff --git a/configs/integration/batch_key.yaml b/configs/integration/batch_key.yaml
@@ -21,6 +21,8 @@ methods:
   bbknn:
   mnc:
   rpca:
+  fastmnn:
+
 
 random_seeds: 5
 n_samples: 
@@ -39,7 +41,4 @@ scenarios:
     labels_path: labels/neftel_ss2.csv
   neftel_10x:
     data_path: data/raw/Neftel_brain_10X.h5ad
-    labels_path: labels/neftel_10x.csv
-  vazquez_ovarian:
-    data_path: data/raw/vazquez_ovarian.h5ad
-    labels_path: labels/vazquez_ovarian.csv
+    labels_path: labels/neftel_10x.csv
diff --git a/configs/integration/batch_key_rms.yaml b/configs/integration/batch_key_rms.yaml
@@ -0,0 +1,46 @@
+root: results/integration/batch_key
+
+plotting: False
+
+n_hvgs:
+  - "2000"
+
+batch_key:
+  - True
+  - False
+
+methods:
+  cca:
+  unintegrated:
+  combat:
+  ldvae:
+  dhaka:
+  harmony:
+  scanorama:
+  scvi:
+  bbknn:
+  mnc:
+  rpca:
+  fastmnn:
+
+
+random_seeds: 5
+n_samples: 
+  - 0.8
+
+
+scenarios:
+  rms_all:
+    data_path: data/raw/rms_all.h5ad
+    mem_scaling: 3.0
+    labels_path: labels/RMS.csv
+
+  rms_scrna:
+    data_path: data/raw/rms_scRNAseq.h5ad
+    mem_scaling: 2.0
+    labels_path: labels/RMS.csv
+
+  rms_snrna:
+    data_path: data/raw/rms_snRNAseq.h5ad
+    mem_scaling: 2.0
+    labels_path: labels/RMS.csv
diff --git a/configs/integration/batch_key_vazquez.yaml b/configs/integration/batch_key_vazquez.yaml
@@ -21,6 +21,7 @@ methods:
   bbknn:
   mnc:
   rpca:
+  fastmnn:
 
 random_seeds: 5
 n_samples: 
diff --git a/configs/integration/n_hvgs.yaml b/configs/integration/n_hvgs.yaml
@@ -13,7 +13,6 @@ batch_key:
   - True
 
 methods:
-  cca:
   unintegrated:
   combat:
   ldvae:
@@ -24,6 +23,8 @@ methods:
   bbknn:
   mnc:
   rpca:
+  fastmnn:
+
 
 random_seeds: 5
 n_samples: 
diff --git a/configs/integration/umap.yaml b/configs/integration/umap.yaml
@@ -20,6 +20,7 @@ methods:
   bbknn:
   mnc:
   rpca:
+  fastmnn:
 
 random_seeds: 1
 
@@ -36,7 +37,4 @@ scenarios:
     labels_path: labels/neftel_ss2.csv
   neftel_10x:
     data_path: data/raw/Neftel_brain_10X.h5ad
-    labels_path: labels/neftel_10x.csv
-  vazquez_ovarian:
-    data_path: data/raw/vazquez_ovarian.h5ad
-    labels_path: labels/vazquez_ovarian.csv
+    labels_path: labels/neftel_10x.csv
diff --git a/configs/integration/umap_rms.yaml b/configs/integration/umap_rms.yaml
@@ -1,4 +1,4 @@
-root: results/integration/umap_rms
+root: results/integration/umap
 
 plotting: True
 skip_preprocessing: True
@@ -11,11 +11,16 @@ batch_key:
 
 methods:
   unintegrated:
+  combat:
+  ldvae:
+  dhaka:
   harmony:
+  scanorama:
   scvi:
   bbknn:
   mnc:
   rpca:
+  fastmnn:
 
 random_seeds: 1
 
diff --git a/configs/integration/umap_wu.yaml b/configs/integration/umap_wu.yaml
@@ -20,6 +20,7 @@ methods:
   bbknn:
   mnc:
   rpca:
+  fastmnn:
 
 random_seeds: 1
 
diff --git a/configs/signatures/breast.yml b/configs/signatures/breast.yml
@@ -16,6 +16,8 @@ methods:
   mnc:
   scalop:
   genenmf:
+  rpca:
+  fastmnn:
 
 random_seeds: 5
 
diff --git a/configs/signatures/debug.yml b/configs/signatures/debug.yml
@@ -9,27 +9,20 @@ batch_key:
 
 methods:
   cca:
-  unintegrated:
-  harmony:
-  bbknn:
-  mnc:
-  scalop:
-
-
-skip_preprocessing: True
+  fastmnn:
 
 random_seeds: 1
 
 scenarios:
-  neftel_0.0_0:
-    data_path: results/signatures/gbm/data/preprocessed/neftel_smartseq2_brain.h5ad
-    mem_scaling: 0.5
+  ji_skin:
+    data_path: data/raw/Ji_skin.h5ad
+    random_seeds: 5
     signatures:
-      - neftel_4
+      - ji_3
 
 signatures:
-  neftel_4:
-    annotation_path: annotations/neftel_4.csv
-    scoring_scenario: neftel_0.0_0
-    n_cluster: 4
+  ji_3:
+    annotation_path: annotations/ji_skin.csv
+    scoring_scenario: ji_skin
+    n_cluster: 3
 
diff --git a/configs/signatures/gbm.yml b/configs/signatures/gbm.yml
@@ -17,6 +17,9 @@ methods:
   mnc:
   scalop:
   genenmf:
+  fastmnn:
+  rpca:
+
 
 random_seeds: 5
 
@@ -63,10 +66,10 @@ scenarios:
 
   neftel_10x_brain:
     data_path: data/raw/Neftel_brain_10X.h5ad
-    preprocessing:
-      min_genes: 1000
-      max_pct_mt: 20
     bulk_path: TCGA_data/GBM/GBM_gex_fpkm_uq.csv
+    preprocessing:
+      excluded_samples:
+        - "114"
     signatures:
       - neftel_4
       - neftel_6
diff --git a/configs/signatures/gbm_subsample.yml b/configs/signatures/gbm_subsample.yml
@@ -16,6 +16,10 @@ methods:
   mnc:
   scalop:
   genenmf:
+  fastmnn:
+  rpca:
+
+
 
 n_samples:
   - 5
diff --git a/configs/signatures/luad.yml b/configs/signatures/luad.yml
@@ -16,6 +16,10 @@ methods:
   mnc:
   scalop:
   genenmf:
+  fastmnn:
+  rpca:
+
+
 
 random_seeds: 5
 
diff --git a/configs/signatures/rms.yml b/configs/signatures/rms.yml
@@ -21,12 +21,14 @@ scenarios:
 
   rms_all:
     data_path: data/raw/rms_all.h5ad
+    mem_scaling: 0.5
     signatures:
       - rms_6
       - rms_scanpy
 
   rms_scrna:
     data_path: data/raw/rms_scRNAseq.h5ad
+    mem_scaling: 0.5
     signatures:
       - rms_6
       - rms_scanpy
diff --git a/configs/signatures/scc.yml b/configs/signatures/scc.yml
@@ -16,6 +16,9 @@ methods:
   mnc:
   scalop:
   genenmf:
+  fastmnn:
+  rpca:
+
 
 
 random_seeds: 5
diff --git a/configs/signatures/sparse.yml b/configs/signatures/sparse.yml
@@ -17,6 +17,10 @@ methods:
   mnc:
   scalop:
   genenmf:
+  fastmnn:
+  rpca:
+
+
 
 
 skip_preprocessing: True
diff --git a/profiles/config.yaml b/profiles/config.yaml
@@ -6,8 +6,3 @@ default-resources:
     - runtime=360
     - partition="compute"
     - slurm_extra=""
-
-group-components:
-  - signature_eval=20
-  - overlap_eval=20
-  - fast_integration=20
diff --git a/rules/evaluate_integration/Snakefile b/rules/evaluate_integration/Snakefile
@@ -5,7 +5,7 @@ rule eval_latent:
         metrics=cfg.ROOT / "eval/{scenario}_{n_samples}_{random_seed}/{n_hvg}/{batch_key}/{method}/metrics.csv",
         umap=cfg.ROOT / "plotting/{scenario}_{n_samples}_{random_seed}/{n_hvg}/{batch_key}/{method}/umap.csv"
     wildcard_constraints:
-        method="scvi.*|harmony.*|unintegrated.*|mnc.*|cca.*|ldvae.*|combat.*|dhaka.*|scanorama.*|cca.*|rpca.*"
+        method="scvi.*|harmony.*|unintegrated.*|mnc.*|cca.*|ldvae.*|combat.*|dhaka.*|scanorama.*|cca.*|rpca.|fastmnn.*"
     container: "docker://fbarkmann/cansig:base"
     threads: 4
     log: cfg.ROOT /  "logs/eval/{method}/{scenario}_{n_samples}_{random_seed}/{n_hvg}/{batch_key}.txt"
diff --git a/rules/integration/Snakefile b/rules/integration/Snakefile
@@ -8,7 +8,7 @@ rule integration_base:
     wildcard_constraints: 
         method="test.*"
     log:
-        cfg.ROOT / "logs/meta_sigs/{scenario}_{n_samples}_{random_seed}/{n_hvg}/{batch_key}/{method}/integration.txt"
+        cfg.ROOT / "logs/integration/{scenario}_{n_samples}_{random_seed}/{n_hvg}/{batch_key}/{method}/integration.txt"
     params: 
         script=""
     shell:
@@ -116,7 +116,7 @@ rule cca_latent:
     benchmark:
         cfg.ROOT / "benchmarks/{scenario}_{n_samples}_{random_seed}/{n_hvg}/{batch_key}/{method}/integration_{ext}.txt"
     log:
-        cfg.ROOT / "logs/meta_sigs/{scenario}_{n_samples}_{random_seed}/{n_hvg}/{batch_key}/{method}/integration_{ext}.txt"
+        cfg.ROOT / "logs/integration/{scenario}_{n_samples}_{random_seed}/{n_hvg}/{batch_key}/{method}/integration_{ext}.txt"
     wildcard_constraints: 
         method="cca.*",
         ext="csv|rds"
@@ -144,3 +144,28 @@ use rule cca_latent as rpca_latent with:
         script="scripts/integration/seurat_integration.R",
         method="rpca"
 
+rule fastmnn_latent:
+    input: 
+        rules.convert_h5ad_to_rds.output
+    output: 
+        cfg.ROOT / "integration/{method}/{scenario}_{n_samples}_{random_seed}/{n_hvg}/{batch_key}/latent.{ext}"
+    container: "docker://fbarkmann/cansig:r"
+    benchmark:
+        cfg.ROOT / "benchmarks/{scenario}_{n_samples}_{random_seed}/{n_hvg}/{batch_key}/{method}/integration_{ext}.txt"
+    log:
+        cfg.ROOT / "logs/integration/{scenario}_{n_samples}_{random_seed}/{n_hvg}/{batch_key}/{method}/integration_{ext}.txt"
+    wildcard_constraints: 
+        method="fastmnn.*",
+        ext="csv|rds"
+    threads: 6
+    resources:
+        mem_mb= lambda x: min(128000, 128000 * cfg.get_mem_scaling(x))
+    params: 
+        script="scripts/integration/fastMNN.R",
+    shell:"""Rscript {params.script} \
+            -i {input} \
+            -o {output} \
+            -r {wildcards.random_seed} \
+            --n-hvg {wildcards.n_hvg} \
+            --batch-key {wildcards.batch_key} \
+            &> {log}"""
diff --git a/rules/signatures/Snakefile b/rules/signatures/Snakefile
@@ -1,7 +1,7 @@
 
 _SINGATURES_OUT = cfg.ROOT / "signatures/{method}/{scenario}_{n_samples}_{random_seed}/{cluster}/{n_hvg}/{batch_key}/signatures.csv"
 _SINGATURES_BENCHMARK = cfg.ROOT / "benchmarks/{scenario}_{n_samples}_{random_seed}/{method}/{cluster}/{n_hvg}/{batch_key}.txt"
-_SINGATURES_LOG =  cfg.ROOT / "logs/meta_sigs/{scenario}_{n_samples}_{random_seed}/{method}/{cluster}/{n_hvg}/{batch_key}.txt"
+_SINGATURES_LOG =  cfg.ROOT / "logs/signatures/{scenario}_{n_samples}_{random_seed}/{method}/{cluster}/{n_hvg}/{batch_key}.txt"
 
 ruleorder: get_signatures_latent > genenmf
 
@@ -16,7 +16,7 @@ rule get_signatures_r:
     output: _SINGATURES_OUT
     benchmark: _SINGATURES_BENCHMARK
     wildcard_constraints: 
-        method="cca.*|rpca*"
+        method="cca.*|rpca*|fastmnn.*"
     threads: 4
     group: "fast_integration"
     resources:
diff --git a/scripts/integration/fastMNN.R b/scripts/integration/fastMNN.R
diff --git a/scripts/signatures/early_integration/get_signatures.R b/scripts/signatures/early_integration/get_signatures.R
diff --git a/submits/debug.sh b/submits/debug.sh
diff --git a/submits/integration.sh b/submits/integration.sh
diff --git a/submits/signatures.sh b/submits/signatures.sh
