visze · Ampholyt · Oct 28, 2024 · Jan 21, 2025 · visze · Jan 21, 2025
diff --git a/config/sbatch.yml b/config/sbatch.yml
@@ -1,48 +1,58 @@
 ---
 __default__:
-  time: "0-00:20"
-  threads: 1
-  mem: 2G
-  queue: debug
-  error: "logs/%x_%j_%N.err"
-  output: "logs/%x_%j_%N.log"
+    time: "0-00:20:00"
+    threads: 1
+    mem: 2G
+    queue: debug
+    error: "logs/%x_%j_%N.err"
+    output: "logs/%x_%j_%N.log"
+regression_input_get_training:
+    time: "0-02:00"
+    threads: 1
+    mem: 5G
+    queue: short
+regression_input_get_test:
+    time: "0-02:00"
+    threads: 1
+    mem: 5G
+    queue: short
 sequence_selection_negative_background_widnows:
-  time: "0-01:00"
-  threads: 1
-  mem: 10G
+    time: "0-01:00:00"
+    threads: 1
+    mem: 10G
 sequence_selection_negative_background_input:
-  time: "0-01:00"
-  threads: 1
-  mem: 10G
+    time: "0-01:00"
+    threads: 1
+    mem: 10G
 sequence_selection_negative_background_sampler:
-  time: "0-01:00"
-  threads: 1
-  mem: 20G
+    time: "0-01:00"
+    threads: 1
+    mem: 20G
 training_multitask:
-  time: "5-00:00"
-  mem: 30G
-  threads: 25
-  queue: gpu --gres=gpu:tesla:4
+    time: "5-00:00"
+    mem: 30G
+    threads: 25
+    queue: gpu --gres=gpu:tesla:2
 predict_fromFasta:
-  time: "0-02:00"
-  mem: 30G
-  threads: 1
-  queue: gpu --gres=gpu:tesla:1
+    time: "0-02:00"
+    mem: 30G
+    threads: 1
+    queue: gpu --gres=gpu:tesla:1
 training_regression:
     time: "0-08:00"
-    mem: 30G
-    threads: 25
-    queue: gpu --gres=gpu:tesla:4
+    mem: 10G
+    threads: 1
+    queue: gpu --gres=gpu:tesla:1
 predict_regression:
     time: "0-02:00"
     mem: 30G
     threads: 1
     queue: gpu --gres=gpu:tesla:1
 model_interpretation_runBackgroundSampler:
-  time: "0-10:00"
-  threads: 4
-  mem: 80G
-  queue: medium
+    time: "0-10:00"
+    threads: 4
+    mem: 80G
+    queue: medium
 model_interpretation_deepexplain:
     time: "1-00:00"
     mem: 30G
@@ -72,4 +82,4 @@ plot_satmut_predict_region:
     time: "0-00:10"
     mem: 10G
     threads: 1
-    queue: gpu --gres=gpu:tesla:1
+    queue: gpu --gres=gpu:tesla:1
diff --git a/workflow/envs/deepexplain_gpu.yml b/workflow/envs/deepexplain_gpu.yml
@@ -1,7 +1,6 @@
 channels:
   - conda-forge
   - bioconda
-  - defaults
 channel_priority: true
 dependencies:
   - python

diff --git a/workflow/envs/deeplift_gpu.yml b/workflow/envs/deeplift_gpu.yml
@@ -1,7 +1,6 @@
 channels:
   - conda-forge
   - bioconda
-  - defaults
 dependencies:
   - click
   - kipoi

diff --git a/workflow/envs/default.yml b/workflow/envs/default.yml
@@ -3,7 +3,6 @@
 channels:
   - conda-forge
   - bioconda
-  - defaults
 dependencies:
   - bc
   - click

diff --git a/workflow/envs/legnet.yml b/workflow/envs/legnet.yml
@@ -3,7 +3,6 @@ name: legnet
 channels:
   - conda-forge
   - bioconda
-  - defaults
 dependencies:
   - _libgcc_mutex=0.1=conda_forge
   - _openmp_mutex=4.5=2_kmp_llvm

diff --git a/workflow/envs/negative_training_sampler.yml b/workflow/envs/negative_training_sampler.yml
@@ -3,6 +3,5 @@
 channels:
   - conda-forge
   - bioconda
-  - defaults
 dependencies:
   - negative_training_sampler
diff --git a/workflow/envs/plot_ism_regions.yaml b/workflow/envs/plot_ism_regions.yaml
@@ -2,7 +2,6 @@
 channels:
   - conda-forge
   - bioconda
-  - defaults
 dependencies:
   - click
   - h5py

diff --git a/workflow/envs/plot_satmut.yaml b/workflow/envs/plot_satmut.yaml
@@ -2,7 +2,6 @@
 channels:
   - conda-forge
   - bioconda
-  - defaults
 dependencies:
   - r-tidyverse
   - r-optparse

diff --git a/workflow/envs/plot_satmut_regions.yaml b/workflow/envs/plot_satmut_regions.yaml
@@ -2,7 +2,6 @@
 channels:
   - conda-forge
   - bioconda
-  - defaults
 dependencies:
   - click
   - h5py

diff --git a/workflow/envs/samtools.yaml b/workflow/envs/samtools.yaml
@@ -3,7 +3,6 @@
 channels:
   - conda-forge
   - bioconda
-  - defaults
 dependencies:
   - samtools
   - bedtools

diff --git a/workflow/envs/tensorflow.yml b/workflow/envs/tensorflow.yml
@@ -2,15 +2,18 @@
 channels:
   - conda-forge
   - bioconda
-  - defaults
+  - anaconda
 dependencies:
-  - click
-  - kipoi
-  - kipoi-utils
-  - kipoiseq
-  - numpy
-  - pandas
-  - python
+  - click==8.1.8
+  - kipoi==0.8.6
+  - kipoi-utils==0.7.7
+  - kipoiseq==0.7.1
+  - numpy==1.23.4
+  - pandas==1.3.5
+  - python==3.9
   - pybedtools
-  - tensorflow
-  - tensorflow-gpu
+  - tensorflow==2.6.2
+  - tensorflow-gpu <2.14
+  - cudatoolkit
+  - cudnn
+  - nccl
diff --git a/workflow/envs/tfmodisco-lite.yml b/workflow/envs/tfmodisco-lite.yml
@@ -2,7 +2,6 @@
 channels:
   - conda-forge
   - bioconda
-  - defaults
 dependencies:
   - click
   - h5py

diff --git a/workflow/envs/tfmodisco.yml b/workflow/envs/tfmodisco.yml
@@ -2,7 +2,6 @@
 channels:
   - conda-forge
   - bioconda
-  - defaults
 dependencies:
   - click
   - h5py

diff --git a/workflow/rules/common.smk b/workflow/rules/common.smk
@@ -1,14 +1,12 @@
 ################################
 #### Global functions       ####
 ################################
-from snakemake.workflow import srcdir
-
-SCRIPTS_DIR = srcdir("../scripts")
-RESOURCES_DIR = srcdir("../../resources")
-
+import os
+RESOURCES_DIR = os.path.abspath("../../resources")
+SCRIPTS_DIR = "../scripts"
 
 def getScript(name):
-    return "%s/%s" % (SCRIPTS_DIR, name)
+    return os.path.join(os.path.dirname(workflow.main_snakefile), 'scripts', name)
 
 
 from snakemake.utils import validate
@@ -56,7 +54,8 @@ def getWrapperPath(file):
     """
     Get directory for snakemake wrappers.
     """
-    return "%s/%s" % (config["wrapper_directory"], file)
+    abs_path = os.path.abspath(os.path.join(config["wrapper_directory"], file))
+    return abs_path
 
 
 def isPretrained():

diff --git a/workflow/rules/model_interpretation.smk b/workflow/rules/model_interpretation.smk
@@ -216,7 +216,7 @@ rule model_interpretation_ism:
 ## tf-modisco-ism
 rule model_interpretation_ism_concat:
     conda:
-        "../envs/tfmodisco.yml"
+        "../envs/tfmodisco-lite.yml"
     input:
         scores=lambda wc: expand(
             "results/model_interpretation/ism/scores.{{test_fold}}.{validation_fold}.h5",
@@ -237,26 +237,26 @@ rule model_interpretation_ism_concat:
         """
 
 
-rule model_interpretation_tfmodisco_ism:
-    conda:
-        "../envs/tfmodisco.yml"
-    input:
-        scores=expand(
-            "results/model_interpretation/ism/scores_concat.{test_fold}.h5",
-            test_fold=list(range(1, 11)),
-        ),
-        script=getScript("tfmodisco_single_task.py"),
-    output:
-        hdf5="results/model_interpretation/ism/tfmodisco/tfmodisco.hdf5",
-    log:
-        "logs/model_interpretation/tfmodisco_ism.log",
-    shell:
-        """
-        scores=`for i in {input.scores}; do echo "--scores $i"; done`;
-        python {input.script} \
-        `echo $scores` \
-        --output {output.hdf5}  &> {log}
-        """
+# rule model_interpretation_tfmodisco_ism:
+#     conda:
+#         "../envs/tfmodisco.yml"
+#     input:
+#         scores=expand(
+#             "results/model_interpretation/ism/scores_concat.{test_fold}.h5",
+#             test_fold=list(range(1, 11)),
+#         ),
+#         script=getScript("tfmodisco_single_task.py"),
+#     output:
+#         hdf5="results/model_interpretation/ism/tfmodisco/tfmodisco.hdf5",
+#     log:
+#         "logs/model_interpretation/tfmodisco_ism.log",
+#     shell:
+#         """
+#         scores=`for i in {input.scores}; do echo "--scores $i"; done`;
+#         python {input.script} \
+#         `echo $scores` \
+#         --output {output.hdf5}  &> {log}
+#         """
 
 
 rule model_interpretation_tfmodisco_lite_ism:
@@ -306,26 +306,26 @@ rule model_interpretation_tfmodisco_lite_report_ism:
 ## tf-modisco1
 
 
-rule model_interpretation_tfmodisco1:
-    conda:
-        "../envs/tfmodisco.yml"
-    input:
-        scores="results/model_interpretation/ism/scores.{test_fold}.{validation_fold}.h5",
-        script=getScript("tf_modisco.py"),
-    output:
-        hdf5="results/model_interpretation/ism/tfmodisco/tfModisco.{test_fold}.{validation_fold}.hdf5",
-        fig_dir=directory(
-            "results/model_interpretation/ism/tfmodisco/figures.{test_fold}.{validation_fold}/"
-        ),
-    log:
-        "logs/model_interpretation/tfmodisco1.{test_fold}.{validation_fold}.log",
-    shell:
-        """
-        python {input.script} \
-        --input {input.scores} \
-        --output {output.hdf5} \
-        --output-figures-dir {output.fig_dir} &> {log}
-        """
+# rule model_interpretation_tfmodisco1:
+#     conda:
+#         "../envs/tfmodisco.yml"
+#     input:
+#         scores="results/model_interpretation/ism/scores.{test_fold}.{validation_fold}.h5",
+#         script=getScript("tf_modisco.py"),
+#     output:
+#         hdf5="results/model_interpretation/ism/tfmodisco/tfModisco.{test_fold}.{validation_fold}.hdf5",
+#         fig_dir=directory(
+#             "results/model_interpretation/ism/tfmodisco/figures.{test_fold}.{validation_fold}/"
+#         ),
+#     log:
+#         "logs/model_interpretation/tfmodisco1.{test_fold}.{validation_fold}.log",
+#     shell:
+#         """
+#         python {input.script} \
+#         --input {input.scores} \
+#         --output {output.hdf5} \
+#         --output-figures-dir {output.fig_dir} &> {log}
+#         """
 
 
 rule model_interpretation_tfmodisco2:

diff --git a/workflow/rules/plot_ism.smk b/workflow/rules/plot_ism.smk
@@ -1,7 +1,7 @@
 wildcard_constraints:
-    region="[^/.]+",
-    sequence_length="\d+",
-    target="\d+",
+    region=r"[^/.]+",
+    sequence_length=r"\d+",
+    target=r"\d+",
 
 
 rule plot_ism_get_region:

diff --git a/workflow/rules/training.smk b/workflow/rules/training.smk
@@ -75,7 +75,6 @@ else:
             else "",
         conda:
             "../envs/tensorflow.yml"
-        threads: 25
         shell:
             """
             python {input.script} \