Upload folder using huggingface_hub

README.md

@@ -0,0 +1,120 @@
+---
+license: mit
+library_name: chrombpnet
+tags:
+  - encode
+  - chrombpnet
+  - chromatin-accessibility
+  - ATAC
+  - GM21360
+  - hg38
+---
+# ENCODE ChromBPNet Atlas
+As part of the ENCODE 4 Project, we trained ChromBPNet models on 1,512 ENCODE DNAse-seq and ATAC-seq across 408 biosamples. Here, we provide all models for open-source use.
+
+For more information about the models, see:
+- Main ENCODE 4 Paper
+- [A unified lexicon of predictive DNA sequence motifs from ENCODE transcription factor binding and chromatin accessibility assays](https://doi.org/10.5281/zenodo.17123347) (Deshpande et al., Zenodo 2025)
+- [ChromBPNet: bias factorized, base-resolution deep learning models of chromatin accessibility reveal cis-regulatory sequence syntax, transcription factor footprints and regulatory variants](https://doi.org/10.1101/2024.12.25.630221) (Pampari et al., bioRxiv 2024)
+
+## ChromBPNet model: ATAC in GM21360 (ENCSR105SCQ)
+- Model: ChromBPNet
+- Assay: ATAC-seq
+- Experiment: [ENCSR105SCQ](https://www.encodeproject.org/experiments/ENCSR105SCQ/)
+- Model annotation: [ENCSR185PHM](https://www.encodeproject.org/annotations/ENCSR185PHM/)
+- Biosample: GM21360 (Full name: Homo sapiens GM21360)
+- Cell slim(s): None
+- Organ slim(s): None
+- Developmental slim(s): None
+- System slim(s): None
+- Assembly: hg38
+
+## Directory structure
+- `fold_0`: Model of 5-fold cross-validation: Fold 0
+    - `model.chrombpnet.fold_0.encid.h5`: full chrombpnet model that combines both bias and corrected model in .h5 format
+    - `model.chrombpnet_nobias.fold_0.encid.h5`: bias-corrected accessibility model in .h5 format (Use for all biological discovery)
+    - `model.bias_scaled.fold_0.encid.h5`: bias model in .h5 format
+    - `model.chrombpnet.fold_0.encid.tar`: full chrombpnet model that combines both bias and  corrected model in SavedModel format. After being untarred, it results in a directory named "chrombpnet".
+    - `model.chrombpnet_nobias.fold_0.encid.tar`: bias-corrected accessibility model in SavedModel format (Use for all biological discovery). After being untarred, it results in a directory named "chrombpnet_wo_bias".
+    - `model.bias_scaled.fold_0.encid.tar`: bias model in SavedModel format. After being untarred, it results in a directory named "bias_model_scaled".
+    - `logs.models.fold_0.encid`: folder containing log files for training models
+- `fold_1`: Model of 5-fold coss-validation: Fold 1
+- `fold_2`: Model of 5-fold cross-validation: Fold 2
+- `fold_3`: Model of 5-fold cross-validation: Fold 3
+- `fold_4`: Model of 5-fold cross-validation: Fold 4
+
+# Instructions
+## 1. Pseudocode for loading models in .h5 format 
+
+(1) Use the code in python after appropriately defining `model_in_h5_format` and `inputs`. \
+(2) `inputs` is a one hot encoded sequence of shape (N,2114,4). Here N corresponds to the 
+number of tested sequences, 2114 is the input sequence length and 4 corresponds to [A,C,G,T].
+
+```python
+import tensorflow as tf
+from tensorflow.keras.utils import get_custom_objects
+from tensorflow.keras.models import load_model
+
+custom_objects={"tf": tf}
+get_custom_objects().update(custom_objects)
+
+model=load_model(model_in_h5_format,compile=False)
+outputs = model(inputs)
+```
+
+The list `outputs` consists of two elements. The first element has a shape of (N, 1000) and
+contains logit predictions for a 1000-base-pair output. The second element, with a shape of
+(N, 1), contains logcount predictions. To transform these predictions into per-base signals, 
+follow the provided pseudo code lines below.
+
+```python
+import numpy as np
+
+def softmax(x, temp=1):
+    norm_x = x - np.mean(x,axis=1, keepdims=True)
+    return np.exp(temp*norm_x)/np.sum(np.exp(temp*norm_x), axis=1, keepdims=True)
+    
+predictions = softmax(outputs[0]) * (np.exp(outputs[1])-1)
+```
+
+## 2. Pseudocode for loading models in .tar format
+
+(1) First untar the directory as follows `tar -xvf model.tar`. \
+(2) Use the code below in python after appropriately defining `model_dir_untared` and `inputs`. \
+(3) `inputs` is a one hot encoded sequence of shape (N,2114,4). Here N corresponds to the number
+of tested sequences, 2114 is the input sequence length and 4 corresponds to ACGT.
+
+Reference: https://www.tensorflow.org/api_docs/python/tf/saved_model/load
+
+```python
+import tensorflow as tf
+
+model = tf.saved_model.load('model_dir_untared')
+outputs = model.signatures['serving_default'](**{'sequence':inputs.astype('float32')})
+```
+
+The variable `outputs` represents a dictionary containing two key-value pairs. The first key
+is `logits_profile_predictions`, holding a value with a shape of (N, 1000). This value corresponds
+to logit predictions for a 1000-base-pair output. The second key, named `logcount_predictions``, 
+is associated with a value of shape (N, 1), representing logcount predictions. To transform these
+predictions into per-base signals, utilize the provided pseudo code lines mentioned below.
+
+```python
+import numpy as np
+def softmax(x, temp=1):
+    norm_x = x - np.mean(x,axis=1, keepdims=True)
+    return np.exp(temp*norm_x)/np.sum(np.exp(temp*norm_x), axis=1, keepdims=True)
+    
+predictions = softmax(outputs["logits_profile_predictions"]) * (np.exp(outputs["logcount_predictions"])-1)
+```
+
+## Docker image to load and use the models
+- https://hub.docker.com/r/kundajelab/chrombpnet-atlas/ (tag:v1)
+
+## Code for ChromBPNet
+- https://github.com/kundajelab/chrombpnet/
+
+# License & citation
+External data users may freely download, analyze and publish results based on any ENCODE data without restrictions.
+
+Released under the [ENCODE data-use policy](https://www.encodeproject.org/about/data-use-policy/). Please cite the ENCODE Project Consortium and the model software: [ChromBPNet](https://github.com/kundajelab/chrombpnet) (Pampari et al., bioRxiv 2024).

fold_0/logs.models.fold_0.ENCSR105SCQ/logfile.modelling.fold_0.ENCSR105SCQ.args.json

@@ -0,0 +1,50 @@
+{
+    "cmd": "pipeline",
+    "genome": "/oak/stanford/groups/akundaje/ziwei75/atac_seq_pipeline/hg38/GRCh38_no_alt_analysis_set_GCA_000001405.15.fasta",
+    "chrom_sizes": "/oak/stanford/groups/akundaje/ziwei75/atac_seq_pipeline/hg38/GRCh38_EBV.chrom.sizes.tsv",
+    "input_bam_file": "/oak/stanford/groups/akundaje/projects/chromatin-atlas-2022/ATAC/ENCSR105SCQ/preprocessing/bigWigs/ENCSR105SCQ.bigWig",
+    "input_fragment_file": null,
+    "input_tagalign_file": null,
+    "output_dir": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_0/model_outputs",
+    "data_type": "ATAC",
+    "peaks": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_0/model_outputs/auxiliary/filtered.peaks.bed",
+    "nonpeaks": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_0/model_outputs/auxiliary/filtered.nonpeaks.bed",
+    "chr_fold_path": "/oak/stanford/groups/akundaje/projects/chromatin-atlas-2022/splits/fold_0.json",
+    "outlier_threshold": 0.9999,
+    "ATAC_ref_path": null,
+    "DNASE_ref_path": null,
+    "num_samples": 10000,
+    "inputlen": 2114,
+    "outputlen": 1000,
+    "seed": 1234,
+    "epochs": 50,
+    "early_stop": 5,
+    "learning_rate": 0.001,
+    "trackables": [
+        "logcount_predictions_loss",
+        "loss",
+        "logits_profile_predictions_loss",
+        "val_logcount_predictions_loss",
+        "val_loss",
+        "val_logits_profile_predictions_loss"
+    ],
+    "architecture_from_file": "/home/users/vhecht/chrombpnet/chrombpnet/chrombpnet/training/models/chrombpnet_with_bias_model.py",
+    "file_prefix": null,
+    "html_prefix": "./",
+    "bsort": false,
+    "tmpdir": null,
+    "no_st": false,
+    "bias_model_path": "/scratch/users/vhecht/chromatin-atlas-2022/reference/ATAC/fold_0/model.bias.fold_0.ENCSR868FGK.h5",
+    "negative_sampling_ratio": 0.1,
+    "filters": 512,
+    "n_dilation_layers": 8,
+    "max_jitter": 500,
+    "batch_size": 64,
+    "output_prefix": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_0/model_outputs/models/chrombpnet",
+    "bigwig": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_0/model_outputs/auxiliary/data_unstranded.bw",
+    "plus_shift": null,
+    "minus_shift": null,
+    "chr": "chr8",
+    "pwm_width": 24,
+    "params": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_0/model_outputs/logs/chrombpnet_model_params.tsv"
+}

fold_0/logs.models.fold_0.ENCSR105SCQ/logfile.modelling.fold_0.ENCSR105SCQ.bias_formatting.stdout.txt

@@ -0,0 +1 @@
+Converting /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_0/model_outputs/models/bias_model_scaled.h5 to /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_0/new_model_format/bias_model_scaled.tar with get_new_tf_model_format.py

fold_0/logs.models.fold_0.ENCSR105SCQ/logfile.modelling.fold_0.ENCSR105SCQ.chrombpnet_data_params.tsv

@@ -0,0 +1,3 @@
+counts_sum_min_thresh	0.0
+counts_sum_max_thresh	2464.29
+trainings_pts_post_thresh	133824

fold_0/logs.models.fold_0.ENCSR105SCQ/logfile.modelling.fold_0.ENCSR105SCQ.chrombpnet_formatting.stdout.txt

@@ -0,0 +1 @@
+Converting /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_0/model_outputs/models/chrombpnet.h5 to /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_0/new_model_format/chrombpnet.tar with get_new_tf_model_format.py

fold_0/logs.models.fold_0.ENCSR105SCQ/logfile.modelling.fold_0.ENCSR105SCQ.chrombpnet_model_params.tsv

@@ -0,0 +1,9 @@
+counts_loss_weight	12.1
+filters	512
+n_dil_layers	8
+bias_model_path	/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_0/model_outputs/models/bias_model_scaled.h5
+inputlen	2114
+outputlen	1000
+max_jitter	500
+chr_fold_path	/oak/stanford/groups/akundaje/projects/chromatin-atlas-2022/splits/fold_0.json
+negative_sampling_ratio	0.1

fold_0/logs.models.fold_0.ENCSR105SCQ/logfile.modelling.fold_0.ENCSR105SCQ.chrombpnet_no_bias_formatting.stdout.txt

@@ -0,0 +1 @@
+Converting /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_0/model_outputs/models/chrombpnet_nobias.h5 to /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_0/new_model_format/chrombpnet_nobias.tar with get_new_tf_model_format.py

fold_0/logs.models.fold_0.ENCSR105SCQ/logfile.modelling.fold_0.ENCSR105SCQ.epoch_loss.csv

@@ -0,0 +1,15 @@
+epoch,logcount_predictions_loss,logits_profile_predictions_loss,loss,val_logcount_predictions_loss,val_logits_profile_predictions_loss,val_loss
+0,2.302119016647339,327.9379577636719,355.79412841796875,0.7925803661346436,338.2075500488281,347.7977294921875
+1,0.7584488987922668,312.69927978515625,321.8768615722656,0.7028773427009583,329.8257141113281,338.33056640625
+2,0.6926157474517822,308.076171875,316.456298828125,0.6790193915367126,327.2816467285156,335.49774169921875
+3,0.6559340357780457,305.2723388671875,313.2096862792969,0.6457452178001404,325.779541015625,333.59307861328125
+4,0.6063268184661865,302.9233093261719,310.2596435546875,0.6256651282310486,325.644287109375,333.21490478515625
+5,0.5810421109199524,301.14910888671875,308.17938232421875,0.6116350293159485,324.0345764160156,331.4353332519531
+6,0.5613860487937927,299.2748718261719,306.0675048828125,0.610621988773346,323.7157897949219,331.10418701171875
+7,0.5310792922973633,297.7200622558594,304.14599609375,0.6143215894699097,323.6483459472656,331.0816650390625
+8,0.5116394758224487,296.0511169433594,302.2423095703125,0.5889062881469727,323.68310546875,330.8089599609375
+9,0.4817698299884796,295.163330078125,300.992431640625,0.5980368256568909,324.4582214355469,331.6944274902344
+10,0.4636251926422119,294.0061950683594,299.61590576171875,0.5952457785606384,324.24932861328125,331.4517517089844
+11,0.43873879313468933,293.395263671875,298.704345703125,0.581154465675354,324.8405456542969,331.87249755859375
+12,0.41721779108047485,292.3247985839844,297.373046875,0.6188138723373413,325.359130859375,332.8468017578125
+13,0.3965517580509186,291.44659423828125,296.2452697753906,0.5940050482749939,325.1839294433594,332.3714599609375

fold_0/model.bias_scaled.fold_0.ENCSR105SCQ.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33881b3340fc0a5988f2ffb9dc09a2df32014ca5156c6c55c33e4cfafd11b08d
+size 2691928

fold_0/model.bias_scaled.fold_0.ENCSR105SCQ.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:11383f2948766555c3961bcafdf6db992a0334ffb4390c8f4859d479fc782cdf
+size 1198080

fold_0/model.chrombpnet.fold_0.ENCSR105SCQ.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c26adfe4ba8d1b9bfd782210d6c1e9ad0fbe18bd624a565f5d157c5dd322b900
+size 77538896

fold_0/model.chrombpnet.fold_0.ENCSR105SCQ.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e89c033c0ac130256223bfbb88897821da5619e6ff916441c66b8ba3e02b9a27
+size 27535360

fold_0/model.chrombpnet_nobias.fold_0.ENCSR105SCQ.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:25d2f9ffd4bed4bc1bb676813ecf3d16f9f6c580f3cdb118e75d75bb45849308
+size 25582648

fold_0/model.chrombpnet_nobias.fold_0.ENCSR105SCQ.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4638ab2a2961b2c67759642246144fad074ced1fdee4f809f0b7969ef7daa910
+size 26060800

fold_1/logs.models.fold_1.ENCSR105SCQ/logfile.modelling.fold_1.ENCSR105SCQ.args.json

@@ -0,0 +1,50 @@
+{
+    "cmd": "pipeline",
+    "genome": "/oak/stanford/groups/akundaje/ziwei75/atac_seq_pipeline/hg38/GRCh38_no_alt_analysis_set_GCA_000001405.15.fasta",
+    "chrom_sizes": "/oak/stanford/groups/akundaje/ziwei75/atac_seq_pipeline/hg38/GRCh38_EBV.chrom.sizes.tsv",
+    "input_bam_file": "/oak/stanford/groups/akundaje/projects/chromatin-atlas-2022/ATAC/ENCSR105SCQ/preprocessing/bigWigs/ENCSR105SCQ.bigWig",
+    "input_fragment_file": null,
+    "input_tagalign_file": null,
+    "output_dir": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_1/model_outputs",
+    "data_type": "ATAC",
+    "peaks": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_1/model_outputs/auxiliary/filtered.peaks.bed",
+    "nonpeaks": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_1/model_outputs/auxiliary/filtered.nonpeaks.bed",
+    "chr_fold_path": "/oak/stanford/groups/akundaje/projects/chromatin-atlas-2022/splits/fold_1.json",
+    "outlier_threshold": 0.9999,
+    "ATAC_ref_path": null,
+    "DNASE_ref_path": null,
+    "num_samples": 10000,
+    "inputlen": 2114,
+    "outputlen": 1000,
+    "seed": 1234,
+    "epochs": 50,
+    "early_stop": 5,
+    "learning_rate": 0.001,
+    "trackables": [
+        "logcount_predictions_loss",
+        "loss",
+        "logits_profile_predictions_loss",
+        "val_logcount_predictions_loss",
+        "val_loss",
+        "val_logits_profile_predictions_loss"
+    ],
+    "architecture_from_file": "/home/users/vhecht/chrombpnet/chrombpnet/chrombpnet/training/models/chrombpnet_with_bias_model.py",
+    "file_prefix": null,
+    "html_prefix": "./",
+    "bsort": false,
+    "tmpdir": null,
+    "no_st": false,
+    "bias_model_path": "/scratch/users/vhecht/chromatin-atlas-2022/reference/ATAC/fold_0/model.bias.fold_0.ENCSR868FGK.h5",
+    "negative_sampling_ratio": 0.1,
+    "filters": 512,
+    "n_dilation_layers": 8,
+    "max_jitter": 500,
+    "batch_size": 64,
+    "output_prefix": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_1/model_outputs/models/chrombpnet",
+    "bigwig": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_1/model_outputs/auxiliary/data_unstranded.bw",
+    "plus_shift": null,
+    "minus_shift": null,
+    "chr": "chr12",
+    "pwm_width": 24,
+    "params": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_1/model_outputs/logs/chrombpnet_model_params.tsv"
+}

fold_1/logs.models.fold_1.ENCSR105SCQ/logfile.modelling.fold_1.ENCSR105SCQ.bias_formatting.stdout.txt

@@ -0,0 +1 @@
+Converting /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_1/model_outputs/models/bias_model_scaled.h5 to /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_1/new_model_format/bias_model_scaled.tar with get_new_tf_model_format.py

fold_1/logs.models.fold_1.ENCSR105SCQ/logfile.modelling.fold_1.ENCSR105SCQ.chrombpnet_data_params.tsv

@@ -0,0 +1,3 @@
+counts_sum_min_thresh	0.0
+counts_sum_max_thresh	2428.72
+trainings_pts_post_thresh	137274

fold_1/logs.models.fold_1.ENCSR105SCQ/logfile.modelling.fold_1.ENCSR105SCQ.chrombpnet_formatting.stdout.txt

@@ -0,0 +1 @@
+Converting /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_1/model_outputs/models/chrombpnet.h5 to /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_1/new_model_format/chrombpnet.tar with get_new_tf_model_format.py

fold_1/logs.models.fold_1.ENCSR105SCQ/logfile.modelling.fold_1.ENCSR105SCQ.chrombpnet_model_params.tsv

@@ -0,0 +1,9 @@
+counts_loss_weight	12.1
+filters	512
+n_dil_layers	8
+bias_model_path	/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_1/model_outputs/models/bias_model_scaled.h5
+inputlen	2114
+outputlen	1000
+max_jitter	500
+chr_fold_path	/oak/stanford/groups/akundaje/projects/chromatin-atlas-2022/splits/fold_1.json
+negative_sampling_ratio	0.1

fold_1/logs.models.fold_1.ENCSR105SCQ/logfile.modelling.fold_1.ENCSR105SCQ.chrombpnet_no_bias_formatting.stdout.txt

@@ -0,0 +1 @@
+Converting /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_1/model_outputs/models/chrombpnet_nobias.h5 to /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_1/new_model_format/chrombpnet_nobias.tar with get_new_tf_model_format.py

fold_1/logs.models.fold_1.ENCSR105SCQ/logfile.modelling.fold_1.ENCSR105SCQ.epoch_loss.csv

@@ -0,0 +1,16 @@
+epoch,logcount_predictions_loss,logits_profile_predictions_loss,loss,val_logcount_predictions_loss,val_logits_profile_predictions_loss,val_loss
+0,3.122945785522461,327.4517822265625,365.239013671875,0.7175904512405396,335.8476257324219,344.5306396484375
+1,0.7796387672424316,313.0270080566406,322.46038818359375,0.8085824251174927,327.3634948730469,337.14715576171875
+2,0.7091411352157593,308.32025146484375,316.9012451171875,0.6899808049201965,326.46160888671875,334.81036376953125
+3,0.6647468209266663,305.1725158691406,313.21600341796875,0.6335970759391785,324.5455627441406,332.2119445800781
+4,0.6314072608947754,303.63763427734375,311.2777099609375,0.6153064370155334,323.4704895019531,330.9156799316406
+5,0.6002985835075378,301.34161376953125,308.6052551269531,0.6122703552246094,323.1646423339844,330.57318115234375
+6,0.5749509930610657,300.0336608886719,306.9900817871094,0.5329983830451965,322.29248046875,328.7419128417969
+7,0.5457547903060913,297.9784240722656,304.5826110839844,0.6510724425315857,322.9836120605469,330.86138916015625
+8,0.522547721862793,296.25238037109375,302.57550048828125,0.5077223777770996,322.55413818359375,328.6973876953125
+9,0.5041512846946716,295.23748779296875,301.3376770019531,0.5102016925811768,322.33636474609375,328.51007080078125
+10,0.48310139775276184,294.0531005859375,299.89874267578125,0.4945288300514221,322.6368408203125,328.62060546875
+11,0.4540764391422272,293.1838073730469,298.6780090332031,0.48886632919311523,322.8401794433594,328.7552490234375
+12,0.43359917402267456,292.67578125,297.92266845703125,0.5896691679954529,323.8255310058594,330.9603576660156
+13,0.41090959310531616,291.4257507324219,296.3980712890625,0.6005422472953796,323.0957946777344,330.3623046875
+14,0.39419519901275635,290.5532531738281,295.32293701171875,0.5064712166786194,323.62725830078125,329.75555419921875

fold_1/model.bias_scaled.fold_1.ENCSR105SCQ.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:277aeeded7339e181d08cad9229f8cd6635dc665fb587b5b43d6831c137ab269
+size 2691928

fold_1/model.bias_scaled.fold_1.ENCSR105SCQ.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e6ecce350b068d26e83bbaf33a0a18da9075c2c951976f56daec2086649d061
+size 1198080

fold_1/model.chrombpnet.fold_1.ENCSR105SCQ.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42732eee65ea610554829fdf5c276407ab1ccb411baf3f599b920a54b3c043bf
+size 77538896

fold_1/model.chrombpnet.fold_1.ENCSR105SCQ.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a990326666f9b819fe7941206644338ece80a853a766a1abaef64b50eb5ce09
+size 27535360

fold_1/model.chrombpnet_nobias.fold_1.ENCSR105SCQ.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:72c2096b80c42e22c9e08a8dc991a3b2214e139ffae567b1bbaf503af2475caf
+size 25582648

fold_1/model.chrombpnet_nobias.fold_1.ENCSR105SCQ.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6143fec97fea88768921233c967e7318624d74132d78da65ccf0a3b1529bb62b
+size 26060800

fold_2/logs.models.fold_2.ENCSR105SCQ/logfile.modelling.fold_2.ENCSR105SCQ.args.json

@@ -0,0 +1,50 @@
+{
+    "cmd": "pipeline",
+    "genome": "/oak/stanford/groups/akundaje/ziwei75/atac_seq_pipeline/hg38/GRCh38_no_alt_analysis_set_GCA_000001405.15.fasta",
+    "chrom_sizes": "/oak/stanford/groups/akundaje/ziwei75/atac_seq_pipeline/hg38/GRCh38_EBV.chrom.sizes.tsv",
+    "input_bam_file": "/oak/stanford/groups/akundaje/projects/chromatin-atlas-2022/ATAC/ENCSR105SCQ/preprocessing/bigWigs/ENCSR105SCQ.bigWig",
+    "input_fragment_file": null,
+    "input_tagalign_file": null,
+    "output_dir": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_2/model_outputs",
+    "data_type": "ATAC",
+    "peaks": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_2/model_outputs/auxiliary/filtered.peaks.bed",
+    "nonpeaks": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_2/model_outputs/auxiliary/filtered.nonpeaks.bed",
+    "chr_fold_path": "/oak/stanford/groups/akundaje/projects/chromatin-atlas-2022/splits/fold_2.json",
+    "outlier_threshold": 0.9999,
+    "ATAC_ref_path": null,
+    "DNASE_ref_path": null,
+    "num_samples": 10000,
+    "inputlen": 2114,
+    "outputlen": 1000,
+    "seed": 1234,
+    "epochs": 50,
+    "early_stop": 5,
+    "learning_rate": 0.001,
+    "trackables": [
+        "logcount_predictions_loss",
+        "loss",
+        "logits_profile_predictions_loss",
+        "val_logcount_predictions_loss",
+        "val_loss",
+        "val_logits_profile_predictions_loss"
+    ],
+    "architecture_from_file": "/home/users/vhecht/chrombpnet/chrombpnet/chrombpnet/training/models/chrombpnet_with_bias_model.py",
+    "file_prefix": null,
+    "html_prefix": "./",
+    "bsort": false,
+    "tmpdir": null,
+    "no_st": false,
+    "bias_model_path": "/scratch/users/vhecht/chromatin-atlas-2022/reference/ATAC/fold_0/model.bias.fold_0.ENCSR868FGK.h5",
+    "negative_sampling_ratio": 0.1,
+    "filters": 512,
+    "n_dilation_layers": 8,
+    "max_jitter": 500,
+    "batch_size": 64,
+    "output_prefix": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_2/model_outputs/models/chrombpnet",
+    "bigwig": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_2/model_outputs/auxiliary/data_unstranded.bw",
+    "plus_shift": null,
+    "minus_shift": null,
+    "chr": "chr22",
+    "pwm_width": 24,
+    "params": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_2/model_outputs/logs/chrombpnet_model_params.tsv"
+}

fold_2/logs.models.fold_2.ENCSR105SCQ/logfile.modelling.fold_2.ENCSR105SCQ.bias_formatting.stdout.txt

@@ -0,0 +1 @@
+Converting /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_2/model_outputs/models/bias_model_scaled.h5 to /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_2/new_model_format/bias_model_scaled.tar with get_new_tf_model_format.py

fold_2/logs.models.fold_2.ENCSR105SCQ/logfile.modelling.fold_2.ENCSR105SCQ.chrombpnet_data_params.tsv

@@ -0,0 +1,3 @@
+counts_sum_min_thresh	0.0
+counts_sum_max_thresh	2429.59
+trainings_pts_post_thresh	140945

fold_2/logs.models.fold_2.ENCSR105SCQ/logfile.modelling.fold_2.ENCSR105SCQ.chrombpnet_formatting.stdout.txt

@@ -0,0 +1 @@
+Converting /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_2/model_outputs/models/chrombpnet.h5 to /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_2/new_model_format/chrombpnet.tar with get_new_tf_model_format.py

fold_2/logs.models.fold_2.ENCSR105SCQ/logfile.modelling.fold_2.ENCSR105SCQ.chrombpnet_model_params.tsv

@@ -0,0 +1,9 @@
+counts_loss_weight	12.1
+filters	512
+n_dil_layers	8
+bias_model_path	/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_2/model_outputs/models/bias_model_scaled.h5
+inputlen	2114
+outputlen	1000
+max_jitter	500
+chr_fold_path	/oak/stanford/groups/akundaje/projects/chromatin-atlas-2022/splits/fold_2.json
+negative_sampling_ratio	0.1

fold_2/logs.models.fold_2.ENCSR105SCQ/logfile.modelling.fold_2.ENCSR105SCQ.chrombpnet_no_bias_formatting.stdout.txt

@@ -0,0 +1 @@
+Converting /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_2/model_outputs/models/chrombpnet_nobias.h5 to /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_2/new_model_format/chrombpnet_nobias.tar with get_new_tf_model_format.py

fold_2/logs.models.fold_2.ENCSR105SCQ/logfile.modelling.fold_2.ENCSR105SCQ.epoch_loss.csv

@@ -0,0 +1,15 @@
+epoch,logcount_predictions_loss,logits_profile_predictions_loss,loss,val_logcount_predictions_loss,val_logits_profile_predictions_loss,val_loss
+0,2.210479259490967,326.3521423339844,353.0990905761719,0.8152256608009338,337.5213928222656,347.3856506347656
+1,0.7470521926879883,311.6716003417969,320.7109375,0.7596765160560608,330.58819580078125,339.7801208496094
+2,0.6837286353111267,308.1412048339844,316.41448974609375,0.6441224813461304,328.16546630859375,335.9593505859375
+3,0.6403308510780334,305.2729187011719,313.02105712890625,0.7109763622283936,326.2019958496094,334.8047790527344
+4,0.6068205237388611,302.7180480957031,310.0610046386719,0.6120291948318481,325.2699279785156,332.675537109375
+5,0.575707733631134,301.0682067871094,308.03411865234375,0.5880399942398071,324.51812744140625,331.6334228515625
+6,0.5550333261489868,299.11468505859375,305.8302001953125,0.5915294885635376,324.74578857421875,331.90338134765625
+7,0.5271279215812683,297.53973388671875,303.9182434082031,0.5671674013137817,324.7595520019531,331.6221923828125
+8,0.5037209987640381,295.9703063964844,302.06512451171875,0.5580423474311829,324.5278015136719,331.2801208496094
+9,0.47702261805534363,294.7151184082031,300.4872741699219,0.6119643449783325,324.4228210449219,331.8276062011719
+10,0.465392142534256,293.8606262207031,299.4916687011719,0.5819709300994873,325.7286682128906,332.7705078125
+11,0.4385700821876526,292.96734619140625,298.2740173339844,0.5746820569038391,324.6486511230469,331.6021728515625
+12,0.4185021221637726,292.27923583984375,297.34283447265625,0.5754173994064331,325.64007568359375,332.60272216796875
+13,0.4029320478439331,291.2797546386719,296.1551818847656,0.6227749586105347,325.6484375,333.1840515136719

fold_2/model.bias_scaled.fold_2.ENCSR105SCQ.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c3b713b9756034e4bcc788346db05b59537d48d0c2659b7d88ae8908c737877
+size 2691928

fold_2/model.bias_scaled.fold_2.ENCSR105SCQ.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b4aa12034336dfb4a1140fc4674c30df6f70163dca21b9f7a16be8cef7174d64
+size 1198080

fold_2/model.chrombpnet.fold_2.ENCSR105SCQ.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:72a5078d6d63fdf79ca8cd3f0b1108c896a7848b6b5202fd2da2bb70ee067c1a
+size 77538896

fold_2/model.chrombpnet.fold_2.ENCSR105SCQ.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f464ada9866a5ffd686ec452211ed05469cfeb697db7ee7ad296a0b8efbf489e
+size 27535360

fold_2/model.chrombpnet_nobias.fold_2.ENCSR105SCQ.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0f177f8fdb1ad9f945d60e364b4587268c88b1730377ac451de0e2ca727954c2
+size 25582648

fold_2/model.chrombpnet_nobias.fold_2.ENCSR105SCQ.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6656f001d0c7f900b36e5fa9e58e10fb9b5dbf0778ec995b6327c30120b1b70a
+size 26060800

fold_3/logs.models.fold_3.ENCSR105SCQ/logfile.modelling.fold_3.ENCSR105SCQ.args.json

@@ -0,0 +1,50 @@
+{
+    "cmd": "pipeline",
+    "genome": "/oak/stanford/groups/akundaje/ziwei75/atac_seq_pipeline/hg38/GRCh38_no_alt_analysis_set_GCA_000001405.15.fasta",
+    "chrom_sizes": "/oak/stanford/groups/akundaje/ziwei75/atac_seq_pipeline/hg38/GRCh38_EBV.chrom.sizes.tsv",
+    "input_bam_file": "/oak/stanford/groups/akundaje/projects/chromatin-atlas-2022/ATAC/ENCSR105SCQ/preprocessing/bigWigs/ENCSR105SCQ.bigWig",
+    "input_fragment_file": null,
+    "input_tagalign_file": null,
+    "output_dir": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_3/model_outputs",
+    "data_type": "ATAC",
+    "peaks": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_3/model_outputs/auxiliary/filtered.peaks.bed",
+    "nonpeaks": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_3/model_outputs/auxiliary/filtered.nonpeaks.bed",
+    "chr_fold_path": "/oak/stanford/groups/akundaje/projects/chromatin-atlas-2022/splits/fold_3.json",
+    "outlier_threshold": 0.9999,
+    "ATAC_ref_path": null,
+    "DNASE_ref_path": null,
+    "num_samples": 10000,
+    "inputlen": 2114,
+    "outputlen": 1000,
+    "seed": 1234,
+    "epochs": 50,
+    "early_stop": 5,
+    "learning_rate": 0.001,
+    "trackables": [
+        "logcount_predictions_loss",
+        "loss",
+        "logits_profile_predictions_loss",
+        "val_logcount_predictions_loss",
+        "val_loss",
+        "val_logits_profile_predictions_loss"
+    ],
+    "architecture_from_file": "/home/users/vhecht/chrombpnet/chrombpnet/chrombpnet/training/models/chrombpnet_with_bias_model.py",
+    "file_prefix": null,
+    "html_prefix": "./",
+    "bsort": false,
+    "tmpdir": null,
+    "no_st": false,
+    "bias_model_path": "/scratch/users/vhecht/chromatin-atlas-2022/reference/ATAC/fold_0/model.bias.fold_0.ENCSR868FGK.h5",
+    "negative_sampling_ratio": 0.1,
+    "filters": 512,
+    "n_dilation_layers": 8,
+    "max_jitter": 500,
+    "batch_size": 64,
+    "output_prefix": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_3/model_outputs/models/chrombpnet",
+    "bigwig": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_3/model_outputs/auxiliary/data_unstranded.bw",
+    "plus_shift": null,
+    "minus_shift": null,
+    "chr": "chr6",
+    "pwm_width": 24,
+    "params": "/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_3/model_outputs/logs/chrombpnet_model_params.tsv"
+}

fold_3/logs.models.fold_3.ENCSR105SCQ/logfile.modelling.fold_3.ENCSR105SCQ.bias_formatting.stdout.txt

@@ -0,0 +1 @@
+Converting /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_3/model_outputs/models/bias_model_scaled.h5 to /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_3/new_model_format/bias_model_scaled.tar with get_new_tf_model_format.py

fold_3/logs.models.fold_3.ENCSR105SCQ/logfile.modelling.fold_3.ENCSR105SCQ.chrombpnet_data_params.tsv

@@ -0,0 +1,3 @@
+counts_sum_min_thresh	0.0
+counts_sum_max_thresh	2473.31
+trainings_pts_post_thresh	138767

fold_3/logs.models.fold_3.ENCSR105SCQ/logfile.modelling.fold_3.ENCSR105SCQ.chrombpnet_formatting.stdout.txt

@@ -0,0 +1 @@
+Converting /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_3/model_outputs/models/chrombpnet.h5 to /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_3/new_model_format/chrombpnet.tar with get_new_tf_model_format.py

fold_3/logs.models.fold_3.ENCSR105SCQ/logfile.modelling.fold_3.ENCSR105SCQ.chrombpnet_model_params.tsv

@@ -0,0 +1,9 @@
+counts_loss_weight	12.2
+filters	512
+n_dil_layers	8
+bias_model_path	/scratch/users/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_3/model_outputs/models/bias_model_scaled.h5
+inputlen	2114
+outputlen	1000
+max_jitter	500
+chr_fold_path	/oak/stanford/groups/akundaje/projects/chromatin-atlas-2022/splits/fold_3.json
+negative_sampling_ratio	0.1

fold_3/logs.models.fold_3.ENCSR105SCQ/logfile.modelling.fold_3.ENCSR105SCQ.chrombpnet_no_bias_formatting.stdout.txt

@@ -0,0 +1 @@
+Converting /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_3/model_outputs/models/chrombpnet_nobias.h5 to /oak/stanford/groups/akundaje/vhecht/chromatin-atlas-2022/ATAC/ENCSR105SCQ/fold_3/new_model_format/chrombpnet_nobias.tar with get_new_tf_model_format.py