Upload folder using huggingface_hub

README.md

@@ -0,0 +1,120 @@
+---
+license: mit
+library_name: chrombpnet
+tags:
+  - encode
+  - chrombpnet
+  - chromatin-accessibility
+  - ATAC
+  - neuron
+  - 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 motor neuron (ENCSR410DWV)
+- Model: ChromBPNet
+- Assay: ATAC-seq
+- Experiment: [ENCSR410DWV](https://www.encodeproject.org/experiments/ENCSR410DWV/)
+- Model annotation: [ENCSR098BDH](https://www.encodeproject.org/annotations/ENCSR098BDH/)
+- Biosample: motor neuron (Full name: Homo sapiens from a donor with amyotrophic lateral sclerosis; motor neuron)
+- Cell slim(s): neural-cell
+- Organ slim(s): spinal-cord,brain
+- Developmental slim(s): ectoderm
+- System slim(s): central-nervous-system
+- 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.ENCSR410DWV/logfile.modelling.fold_0.ENCSR410DWV.args.json

@@ -0,0 +1,50 @@
+{
+    "cmd": "pipeline",
+    "genome": "/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/hg38.fa",
+    "chrom_sizes": "/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/hg38.chrom.sizes",
+    "input_bam_file": "/oak/stanford/groups/akundaje/projects/chromatin-atlas-2022/ATAC/ENCSR410DWV/preprocessing/bigWigs/ENCSR410DWV.bigWig",
+    "input_fragment_file": null,
+    "input_tagalign_file": null,
+    "output_dir": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_0/",
+    "data_type": "ATAC",
+    "peaks": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_0/auxiliary/filtered.peaks.bed",
+    "nonpeaks": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_0/auxiliary/filtered.nonpeaks.bed",
+    "chr_fold_path": "/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/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/cmyun/chrombpnet/chrombpnet/training/models/chrombpnet_with_bias_model.py",
+    "file_prefix": null,
+    "html_prefix": "./",
+    "bsort": false,
+    "tmpdir": null,
+    "no_st": false,
+    "bias_model_path": "/oak/stanford/groups/akundaje/cmyun/chromatlas/bias_model/ENCFF881REW/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": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_0/models/chrombpnet",
+    "bigwig": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_0/auxiliary/data_unstranded.bw",
+    "plus_shift": null,
+    "minus_shift": null,
+    "chr": "chr8",
+    "pwm_width": 24,
+    "params": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_0/logs/chrombpnet_model_params.tsv"
+}

fold_0/logs.models.fold_0.ENCSR410DWV/logfile.modelling.fold_0.ENCSR410DWV.chrombpnet_data_params.tsv

@@ -0,0 +1,3 @@
+counts_sum_min_thresh	0.0
+counts_sum_max_thresh	22932.97
+trainings_pts_post_thresh	243651

fold_0/logs.models.fold_0.ENCSR410DWV/logfile.modelling.fold_0.ENCSR410DWV.chrombpnet_model_params.tsv

@@ -0,0 +1,9 @@
+counts_loss_weight	85.6
+filters	512
+n_dil_layers	8
+bias_model_path	/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_0/models/bias_model_scaled.h5
+inputlen	2114
+outputlen	1000
+max_jitter	500
+chr_fold_path	/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/splits/fold_0.json
+negative_sampling_ratio	0.1

fold_0/logs.models.fold_0.ENCSR410DWV/logfile.modelling.fold_0.ENCSR410DWV.epoch_loss.csv

@@ -0,0 +1,21 @@
+epoch,logcount_predictions_loss,logits_profile_predictions_loss,loss,val_logcount_predictions_loss,val_logits_profile_predictions_loss,val_loss
+0,3.1479382514953613,1044.6121826171875,1314.0767822265625,1.034663200378418,968.9312133789062,1057.498046875
+1,1.0549201965332031,924.7650146484375,1015.06689453125,0.9614323973655701,936.6226196289062,1018.921630859375
+2,0.9336115717887878,890.8167114257812,970.7302856445312,0.838935136795044,916.2326049804688,988.0455322265625
+3,0.8536310195922852,870.4190673828125,943.4901733398438,0.8518881797790527,908.5701293945312,981.4918212890625
+4,0.7965283393859863,855.4117431640625,923.5950927734375,0.7724440693855286,904.7188720703125,970.8404541015625
+5,0.7699877619743347,843.4400634765625,909.3527221679688,0.943589448928833,904.29736328125,985.0687866210938
+6,0.7331543564796448,833.4353637695312,896.1926879882812,0.7689463496208191,905.3001708984375,971.1226806640625
+7,0.7099965810775757,827.2218627929688,887.9972534179688,0.7160753607749939,905.8223876953125,967.1179809570312
+8,0.6817929148674011,822.083740234375,880.4461059570312,0.7499365210533142,904.4540405273438,968.64892578125
+9,0.6588547825813293,816.8921508789062,873.2908325195312,0.7188165187835693,907.2593994140625,968.790771484375
+10,0.6401948928833008,814.0786743164062,868.8795166015625,0.7716817259788513,908.9888916015625,975.0443725585938
+11,0.6293865442276001,809.8638305664062,863.7382202148438,0.7283236384391785,902.5293579101562,964.874755859375
+12,0.6096192598342896,808.4347534179688,860.6177368164062,0.7829275131225586,902.934814453125,969.95361328125
+13,0.6006826758384705,806.387451171875,857.8060302734375,0.7809509634971619,909.4754638671875,976.3245849609375
+14,0.5776863098144531,804.1549072265625,853.6049194335938,0.7143959403038025,902.5397338867188,963.6924438476562
+15,0.5656338930130005,801.5863037109375,850.0068969726562,0.8052111864089966,903.609619140625,972.5358276367188
+16,0.5572560429573059,801.3062133789062,849.00830078125,0.7182430624961853,903.8555908203125,965.337646484375
+17,0.5511627197265625,799.5157470703125,846.696044921875,0.7194530963897705,904.1636352539062,965.7484741210938
+18,0.5324504375457764,797.8685913085938,843.447021484375,0.7086991667747498,906.0185546875,966.6832885742188
+19,0.5280552506446838,796.3851928710938,841.5875244140625,0.7215370535850525,906.4456176757812,968.209228515625

fold_0/model.bias_scaled.fold_0.ENCSR410DWV.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:607a28de645880ea6aa576d87034d28b6a39e0131d377a0fd04bec180b52c2fe
+size 2691928

fold_0/model.bias_scaled.fold_0.ENCSR410DWV.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1eb4c6b6978137f4f1b4ebc75267af14b3032957d31b8157c7df3c8554928c1
+size 1198080

fold_0/model.chrombpnet.fold_0.ENCSR410DWV.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a6c7056c08dccf14c205373af855b93e97ea92eac838777fd3281babcc7744b8
+size 77538864

fold_0/model.chrombpnet.fold_0.ENCSR410DWV.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ce1c1d7bad1f4377affb66c75c47155f5f4fe1ffd0fc4c7b819e594c07eab609
+size 27535360

fold_0/model.chrombpnet_nobias.fold_0.ENCSR410DWV.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c4a6970f7ef9bea3def2db5ef3dc39d70bbef0ed18a70fc717bf66306bb262b
+size 25582648

fold_0/model.chrombpnet_nobias.fold_0.ENCSR410DWV.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:95513784ee06da422deb20ab400af24414b35c82b9d2914bdd38e7709b0fed9f
+size 26060800

fold_1/logs.models.fold_1.ENCSR410DWV/logfile.modelling.fold_1.ENCSR410DWV.args.json

@@ -0,0 +1,50 @@
+{
+    "cmd": "pipeline",
+    "genome": "/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/hg38.fa",
+    "chrom_sizes": "/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/hg38.chrom.sizes",
+    "input_bam_file": "/oak/stanford/groups/akundaje/projects/chromatin-atlas-2022/ATAC/ENCSR410DWV/preprocessing/bigWigs/ENCSR410DWV.bigWig",
+    "input_fragment_file": null,
+    "input_tagalign_file": null,
+    "output_dir": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_1/",
+    "data_type": "ATAC",
+    "peaks": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_1/auxiliary/filtered.peaks.bed",
+    "nonpeaks": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_1/auxiliary/filtered.nonpeaks.bed",
+    "chr_fold_path": "/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/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/cmyun/chrombpnet/chrombpnet/training/models/chrombpnet_with_bias_model.py",
+    "file_prefix": null,
+    "html_prefix": "./",
+    "bsort": false,
+    "tmpdir": null,
+    "no_st": false,
+    "bias_model_path": "/oak/stanford/groups/akundaje/cmyun/chromatlas/bias_model/ENCFF881REW/fold_1/model.bias.fold_1.ENCSR868FGK.h5",
+    "negative_sampling_ratio": 0.1,
+    "filters": 512,
+    "n_dilation_layers": 8,
+    "max_jitter": 500,
+    "batch_size": 64,
+    "output_prefix": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_1/models/chrombpnet",
+    "bigwig": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_1/auxiliary/data_unstranded.bw",
+    "plus_shift": null,
+    "minus_shift": null,
+    "chr": "chr12",
+    "pwm_width": 24,
+    "params": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_1/logs/chrombpnet_model_params.tsv"
+}

fold_1/logs.models.fold_1.ENCSR410DWV/logfile.modelling.fold_1.ENCSR410DWV.chrombpnet_data_params.tsv

@@ -0,0 +1,3 @@
+counts_sum_min_thresh	0.0
+counts_sum_max_thresh	23276.97
+trainings_pts_post_thresh	248830

fold_1/logs.models.fold_1.ENCSR410DWV/logfile.modelling.fold_1.ENCSR410DWV.chrombpnet_model_params.tsv

@@ -0,0 +1,9 @@
+counts_loss_weight	85.5
+filters	512
+n_dil_layers	8
+bias_model_path	/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_1/models/bias_model_scaled.h5
+inputlen	2114
+outputlen	1000
+max_jitter	500
+chr_fold_path	/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/splits/fold_1.json
+negative_sampling_ratio	0.1

fold_1/logs.models.fold_1.ENCSR410DWV/logfile.modelling.fold_1.ENCSR410DWV.epoch_loss.csv

@@ -0,0 +1,14 @@
+epoch,logcount_predictions_loss,logits_profile_predictions_loss,loss,val_logcount_predictions_loss,val_logits_profile_predictions_loss,val_loss
+0,2.203944206237793,996.1837158203125,1184.62109375,1.3107318878173828,1044.779541015625,1156.8466796875
+1,1.0599137544631958,874.0689697265625,964.693359375,0.9371376037597656,1000.6785888671875,1080.8038330078125
+2,0.9357422590255737,841.4035034179688,921.409423828125,0.9534404873847961,990.8845825195312,1072.40380859375
+3,0.8440616130828857,822.164306640625,894.3323364257812,0.983478307723999,965.060791015625,1049.1480712890625
+4,0.7981798052787781,808.8585815429688,877.1029052734375,0.7662743330001831,958.9470825195312,1024.463134765625
+5,0.7469395399093628,798.3795166015625,862.2420654296875,0.777300238609314,970.843994140625,1037.30322265625
+6,0.7196387052536011,790.5194091796875,852.0487670898438,0.7302762269973755,961.28955078125,1023.7282104492188
+7,0.6906212568283081,784.984130859375,844.0326538085938,0.7082449793815613,951.7726440429688,1012.32763671875
+8,0.6616798639297485,778.6289672851562,835.203125,0.721697986125946,952.3245849609375,1014.029296875
+9,0.6447734832763672,776.1478271484375,831.2749633789062,0.7331726551055908,954.729736328125,1017.4158935546875
+10,0.6276258826255798,772.8797607421875,826.54052734375,0.7257129549980164,952.3231201171875,1014.3720703125
+11,0.6002298593521118,769.413818359375,820.7332763671875,0.7062380909919739,953.706298828125,1014.089599609375
+12,0.5827230215072632,767.9713745117188,817.7952270507812,0.6971551775932312,958.5783081054688,1018.1857299804688

fold_1/model.bias_scaled.fold_1.ENCSR410DWV.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:508579bfc39854d186ed4dbf7e270a55cb63ee45f2496b97122b6a6a455aae92
+size 2691928

fold_1/model.bias_scaled.fold_1.ENCSR410DWV.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec5df5d641fcf6b612621c9a5fbd59f398cbd74c8731b9b6a6967fd3d9ba2f75
+size 1198080

fold_1/model.chrombpnet.fold_1.ENCSR410DWV.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a1aa744037d7641fc86b349b0a703251b1d23dce3062fb7a63238e514e3cc30a
+size 77538808

fold_1/model.chrombpnet.fold_1.ENCSR410DWV.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:04cfb61af3de1482070fb9ad1fb6d3c62ee53a7ddbb3d16552a2ae33a0e441c9
+size 27525120

fold_1/model.chrombpnet_nobias.fold_1.ENCSR410DWV.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:27ec66907007608c6f2370624f775c147ced6341d33f1d9a1e15efcc3df50552
+size 25582648

fold_1/model.chrombpnet_nobias.fold_1.ENCSR410DWV.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d37c2ac080a4e4269f2cdb5f7103ce7e014e9dd6efe6f46e762e7e57176b956
+size 26060800

fold_2/logs.models.fold_2.ENCSR410DWV/logfile.modelling.fold_2.ENCSR410DWV.args.json

@@ -0,0 +1,50 @@
+{
+    "cmd": "pipeline",
+    "genome": "/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/hg38.fa",
+    "chrom_sizes": "/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/hg38.chrom.sizes",
+    "input_bam_file": "/oak/stanford/groups/akundaje/projects/chromatin-atlas-2022/ATAC/ENCSR410DWV/preprocessing/bigWigs/ENCSR410DWV.bigWig",
+    "input_fragment_file": null,
+    "input_tagalign_file": null,
+    "output_dir": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_2/",
+    "data_type": "ATAC",
+    "peaks": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_2/auxiliary/filtered.peaks.bed",
+    "nonpeaks": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_2/auxiliary/filtered.nonpeaks.bed",
+    "chr_fold_path": "/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/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/cmyun/chrombpnet/chrombpnet/training/models/chrombpnet_with_bias_model.py",
+    "file_prefix": null,
+    "html_prefix": "./",
+    "bsort": false,
+    "tmpdir": null,
+    "no_st": false,
+    "bias_model_path": "/oak/stanford/groups/akundaje/cmyun/chromatlas/bias_model/ENCFF881REW/fold_2/model.bias.fold_2.ENCSR868FGK.h5",
+    "negative_sampling_ratio": 0.1,
+    "filters": 512,
+    "n_dilation_layers": 8,
+    "max_jitter": 500,
+    "batch_size": 64,
+    "output_prefix": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_2/models/chrombpnet",
+    "bigwig": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_2/auxiliary/data_unstranded.bw",
+    "plus_shift": null,
+    "minus_shift": null,
+    "chr": "chr22",
+    "pwm_width": 24,
+    "params": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_2/logs/chrombpnet_model_params.tsv"
+}

fold_2/logs.models.fold_2.ENCSR410DWV/logfile.modelling.fold_2.ENCSR410DWV.chrombpnet_data_params.tsv

@@ -0,0 +1,3 @@
+counts_sum_min_thresh	0.0
+counts_sum_max_thresh	23691.93
+trainings_pts_post_thresh	255081

fold_2/logs.models.fold_2.ENCSR410DWV/logfile.modelling.fold_2.ENCSR410DWV.chrombpnet_model_params.tsv

@@ -0,0 +1,9 @@
+counts_loss_weight	85.8
+filters	512
+n_dil_layers	8
+bias_model_path	/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_2/models/bias_model_scaled.h5
+inputlen	2114
+outputlen	1000
+max_jitter	500
+chr_fold_path	/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/splits/fold_2.json
+negative_sampling_ratio	0.1

fold_2/logs.models.fold_2.ENCSR410DWV/logfile.modelling.fold_2.ENCSR410DWV.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
+0,5.060719013214111,1042.3536376953125,1476.5655517578125,1.1809552907943726,1002.1784057617188,1103.504638671875
+1,1.1330556869506836,905.0186767578125,1002.2339477539062,1.0150519609451294,954.3045043945312,1041.3961181640625
+2,0.99052894115448,864.1248168945312,949.1124267578125,0.9349614977836609,936.228759765625,1016.4481811523438
+3,0.8944885730743408,841.4276123046875,918.1744995117188,0.8312685489654541,915.576416015625,986.8992919921875
+4,0.8326607942581177,825.2691040039062,896.7117919921875,0.862922728061676,905.3380737304688,979.3764038085938
+5,0.7751821875572205,813.358642578125,879.8702392578125,0.8191050291061401,896.884033203125,967.1630859375
+6,0.7417919039726257,803.0496215820312,866.6961669921875,0.761161744594574,895.2047729492188,960.5118408203125
+7,0.7085431814193726,795.6412353515625,856.433349609375,0.7365481853485107,893.4829711914062,956.6792602539062
+8,0.6836225390434265,789.2531127929688,847.90625,0.7395067811012268,890.9627075195312,954.4124145507812
+9,0.6611850261688232,784.372802734375,841.1016235351562,0.7619611024856567,891.3787841796875,956.7552490234375
+10,0.6327029466629028,780.6455688476562,834.9328002929688,0.7400492429733276,896.2538452148438,959.7499389648438
+11,0.6154577732086182,777.661376953125,830.4674072265625,0.7515643239021301,895.8280029296875,960.3118896484375
+12,0.6013035774230957,775.5665893554688,827.1589965820312,0.7456300258636475,894.1461791992188,958.120849609375
+13,0.5813829302787781,772.2224731445312,822.1063842773438,0.7678282260894775,896.5513305664062,962.4314575195312

fold_2/model.bias_scaled.fold_2.ENCSR410DWV.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:25138bc3476cf0f8a2201f1b7a48395b83b343128684e3eb8efea82d5d31d5e5
+size 2691928

fold_2/model.bias_scaled.fold_2.ENCSR410DWV.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eba6c1ea0b89074e712d29b1b3bcd4717b3fb7dd5b81a0453142e1bd5ef07895
+size 1198080

fold_2/model.chrombpnet.fold_2.ENCSR410DWV.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13e64e12f88f5919f76588720e24eb45bdb0c8176a015248164a8b9b9f10e22e
+size 77538808

fold_2/model.chrombpnet.fold_2.ENCSR410DWV.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f81bd5c7891119b2b3f4fe21e295de377a41fb968a544910ce3e0f9160a54a3d
+size 27525120

fold_2/model.chrombpnet_nobias.fold_2.ENCSR410DWV.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:382353b83722e9ac7f233b6535be16dd7ff5eca824a4b2e60901d06040f6be0a
+size 25582648

fold_2/model.chrombpnet_nobias.fold_2.ENCSR410DWV.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22c9c6833aef14d4c1a4ec07cb6a2024b246d22f383ea6813598dd4654752b1b
+size 26060800

fold_3/logs.models.fold_3.ENCSR410DWV/logfile.modelling.fold_3.ENCSR410DWV.args.json

@@ -0,0 +1,50 @@
+{
+    "cmd": "pipeline",
+    "genome": "/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/hg38.fa",
+    "chrom_sizes": "/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/hg38.chrom.sizes",
+    "input_bam_file": "/oak/stanford/groups/akundaje/projects/chromatin-atlas-2022/ATAC/ENCSR410DWV/preprocessing/bigWigs/ENCSR410DWV.bigWig",
+    "input_fragment_file": null,
+    "input_tagalign_file": null,
+    "output_dir": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_3/",
+    "data_type": "ATAC",
+    "peaks": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_3/auxiliary/filtered.peaks.bed",
+    "nonpeaks": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_3/auxiliary/filtered.nonpeaks.bed",
+    "chr_fold_path": "/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/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/cmyun/chrombpnet/chrombpnet/training/models/chrombpnet_with_bias_model.py",
+    "file_prefix": null,
+    "html_prefix": "./",
+    "bsort": false,
+    "tmpdir": null,
+    "no_st": false,
+    "bias_model_path": "/oak/stanford/groups/akundaje/cmyun/chromatlas/bias_model/ENCFF881REW/fold_3/model.bias.fold_3.ENCSR868FGK.h5",
+    "negative_sampling_ratio": 0.1,
+    "filters": 512,
+    "n_dilation_layers": 8,
+    "max_jitter": 500,
+    "batch_size": 64,
+    "output_prefix": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_3/models/chrombpnet",
+    "bigwig": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_3/auxiliary/data_unstranded.bw",
+    "plus_shift": null,
+    "minus_shift": null,
+    "chr": "chr6",
+    "pwm_width": 24,
+    "params": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_3/logs/chrombpnet_model_params.tsv"
+}

fold_3/logs.models.fold_3.ENCSR410DWV/logfile.modelling.fold_3.ENCSR410DWV.chrombpnet_data_params.tsv

@@ -0,0 +1,3 @@
+counts_sum_min_thresh	0.0
+counts_sum_max_thresh	23681.56
+trainings_pts_post_thresh	248915

fold_3/logs.models.fold_3.ENCSR410DWV/logfile.modelling.fold_3.ENCSR410DWV.chrombpnet_model_params.tsv

@@ -0,0 +1,9 @@
+counts_loss_weight	85.6
+filters	512
+n_dil_layers	8
+bias_model_path	/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_3/models/bias_model_scaled.h5
+inputlen	2114
+outputlen	1000
+max_jitter	500
+chr_fold_path	/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/splits/fold_3.json
+negative_sampling_ratio	0.1

fold_3/logs.models.fold_3.ENCSR410DWV/logfile.modelling.fold_3.ENCSR410DWV.epoch_loss.csv

@@ -0,0 +1,23 @@
+epoch,logcount_predictions_loss,logits_profile_predictions_loss,loss,val_logcount_predictions_loss,val_logits_profile_predictions_loss,val_loss
+0,2.792313814163208,1008.318603515625,1247.3389892578125,1.1111085414886475,945.6076049804688,1040.718505859375
+1,1.060050368309021,886.0775146484375,976.818603515625,1.0541930198669434,900.2244262695312,990.4629516601562
+2,0.9300038814544678,851.3956298828125,931.0045166015625,1.0035959482192993,878.3263549804688,964.2340698242188
+3,0.8470818400382996,830.8131103515625,903.3236083984375,0.837448239326477,870.222412109375,941.9078979492188
+4,0.7915257215499878,818.0314331054688,885.7847900390625,0.786384105682373,867.3631591796875,934.6773681640625
+5,0.7479575276374817,808.1812133789062,872.2066040039062,0.7606648206710815,863.21728515625,928.3299560546875
+6,0.713771402835846,801.0140380859375,862.1116333007812,0.7697365880012512,860.5785522460938,926.4677124023438
+7,0.6873751878738403,795.5082397460938,854.3464965820312,0.7468382120132446,861.7943115234375,925.7237548828125
+8,0.6648452877998352,790.4458618164062,847.35791015625,0.7398837208747864,861.9120483398438,925.2459716796875
+9,0.6402920484542847,786.2002563476562,841.0098266601562,0.9104639291763306,864.8804931640625,942.8158569335938
+10,0.6198247671127319,784.1627807617188,837.2215576171875,0.7460675835609436,862.8003540039062,926.6634521484375
+11,0.6041611433029175,781.3108520507812,833.0281372070312,0.769572913646698,864.4788208007812,930.3545532226562
+12,0.5823327898979187,778.4684448242188,828.31640625,0.7450774312019348,860.880126953125,924.658935546875
+13,0.5695335268974304,777.3741455078125,826.1259765625,0.7467970848083496,859.1257934570312,923.0513305664062
+14,0.5559858083724976,775.9813232421875,823.5735473632812,0.7650761008262634,866.9741821289062,932.4649047851562
+15,0.540423572063446,774.22314453125,820.4817504882812,0.8129595518112183,861.113525390625,930.703369140625
+16,0.5259806513786316,772.8692626953125,817.8939208984375,0.7360764741897583,858.7386474609375,921.7470092773438
+17,0.518766462802887,771.4850463867188,815.8915405273438,0.7984395027160645,863.95849609375,932.3045043945312
+18,0.5104382038116455,771.0863037109375,814.7777099609375,0.7716714143753052,862.87939453125,928.9348754882812
+19,0.5014333128929138,769.671630859375,812.592529296875,0.7364746332168579,859.5607299804688,922.602783203125
+20,0.49545034766197205,768.45556640625,810.8672485351562,0.7491710782051086,862.2573852539062,926.3860473632812
+21,0.48222658038139343,768.0906372070312,809.37109375,0.7883169054985046,864.6580810546875,932.1368408203125

fold_3/model.bias_scaled.fold_3.ENCSR410DWV.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:862dcd521990c3f8aa5ae4bce733ad0fc9020a97b054a32b25b7dbd624cc1e25
+size 2691928

fold_3/model.bias_scaled.fold_3.ENCSR410DWV.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:250161aec301ff143ad3620ecaa0cbebe208f9c2d4f94b75fc5ba1b56b92313c
+size 1198080

fold_3/model.chrombpnet.fold_3.ENCSR410DWV.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:393d9a5e2d14e5c8919dfb44b8be6cd2734c48bff5d7b0295b994c87b29d4294
+size 77538808

fold_3/model.chrombpnet.fold_3.ENCSR410DWV.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bcc2dbaa5bbf0bcf2f3f88a7aad0eab0eda99784363b79529b51db303da56be3
+size 27525120

fold_3/model.chrombpnet_nobias.fold_3.ENCSR410DWV.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0a5f1af5048c569689a4cacb4aa500defc6ce12fbf7a3a46508e5b0d2b692159
+size 25582648

fold_3/model.chrombpnet_nobias.fold_3.ENCSR410DWV.tar

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:571413be2212c4b56d2f8fb458bc4d9dcfb03a611e213ed96d966ff763f9999d
+size 26060800

fold_4/logs.models.fold_4.ENCSR410DWV/logfile.modelling.fold_4.ENCSR410DWV.args.json

@@ -0,0 +1,50 @@
+{
+    "cmd": "pipeline",
+    "genome": "/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/hg38.fa",
+    "chrom_sizes": "/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/hg38.chrom.sizes",
+    "input_bam_file": "/oak/stanford/groups/akundaje/projects/chromatin-atlas-2022/ATAC/ENCSR410DWV/preprocessing/bigWigs/ENCSR410DWV.bigWig",
+    "input_fragment_file": null,
+    "input_tagalign_file": null,
+    "output_dir": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_4/",
+    "data_type": "ATAC",
+    "peaks": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_4/auxiliary/filtered.peaks.bed",
+    "nonpeaks": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_4/auxiliary/filtered.nonpeaks.bed",
+    "chr_fold_path": "/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/splits/fold_4.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/cmyun/chrombpnet/chrombpnet/training/models/chrombpnet_with_bias_model.py",
+    "file_prefix": null,
+    "html_prefix": "./",
+    "bsort": false,
+    "tmpdir": null,
+    "no_st": false,
+    "bias_model_path": "/oak/stanford/groups/akundaje/cmyun/chromatlas/bias_model/ENCFF881REW/fold_4/model.bias.fold_4.ENCSR868FGK.h5",
+    "negative_sampling_ratio": 0.1,
+    "filters": 512,
+    "n_dilation_layers": 8,
+    "max_jitter": 500,
+    "batch_size": 64,
+    "output_prefix": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_4/models/chrombpnet",
+    "bigwig": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_4/auxiliary/data_unstranded.bw",
+    "plus_shift": null,
+    "minus_shift": null,
+    "chr": "chr10",
+    "pwm_width": 24,
+    "params": "/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_4/logs/chrombpnet_model_params.tsv"
+}

fold_4/logs.models.fold_4.ENCSR410DWV/logfile.modelling.fold_4.ENCSR410DWV.chrombpnet_data_params.tsv

@@ -0,0 +1,3 @@
+counts_sum_min_thresh	0.0
+counts_sum_max_thresh	22621.64
+trainings_pts_post_thresh	255186

fold_4/logs.models.fold_4.ENCSR410DWV/logfile.modelling.fold_4.ENCSR410DWV.chrombpnet_model_params.tsv

@@ -0,0 +1,9 @@
+counts_loss_weight	83.9
+filters	512
+n_dil_layers	8
+bias_model_path	/oak/stanford/groups/akundaje/cmyun/chromatlas/ATAC/ENCSR410DWV/chrombpnet_model_fold_4/models/bias_model_scaled.h5
+inputlen	2114
+outputlen	1000
+max_jitter	500
+chr_fold_path	/oak/stanford/groups/akundaje/cmyun/chromatlas/reference/splits/fold_4.json
+negative_sampling_ratio	0.1

fold_4/logs.models.fold_4.ENCSR410DWV/logfile.modelling.fold_4.ENCSR410DWV.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
+0,2.4045910835266113,950.9381713867188,1152.6829833984375,1.0087836980819702,893.81201171875,978.4488525390625
+1,0.9584280252456665,847.8148803710938,928.2258911132812,0.9336455464363098,866.7254028320312,945.0580444335938
+2,0.858567476272583,819.4251708984375,891.4586791992188,0.8298581838607788,851.2421264648438,920.8673095703125
+3,0.7915372848510742,803.6948852539062,870.1052856445312,1.0882701873779297,851.3740844726562,942.6797485351562
+4,0.7538486123085022,793.2411499023438,856.487548828125,0.8198124170303345,855.1895141601562,923.9715576171875
+5,0.7097246646881104,786.3593139648438,845.9049682617188,0.7975136041641235,842.685791015625,909.5967407226562
+6,0.6742368936538696,780.2608032226562,836.830810546875,0.7629029750823975,841.7520141601562,905.7589721679688
+7,0.6373651027679443,775.3582153320312,828.8318481445312,0.9330076575279236,839.050048828125,917.3290405273438
+8,0.6208879947662354,772.5145263671875,824.6082763671875,0.7141630053520203,840.7493896484375,900.6676025390625
+9,0.606909453868866,769.0291748046875,819.9493408203125,0.734984278678894,841.6358642578125,903.300537109375
+10,0.5841914415359497,766.182861328125,815.1964721679688,0.8392884731292725,841.7539672851562,912.170166015625
+11,0.5662336945533752,765.3118286132812,812.8180541992188,0.7625426650047302,839.6192626953125,903.5963134765625
+12,0.5530927181243896,763.042724609375,809.4463500976562,0.8056031465530396,844.3340454101562,911.9240112304688
+13,0.536644458770752,761.8652954101562,806.8895263671875,0.8241577744483948,842.1159057617188,911.2626342773438