add _new_datasets (#2)

- git lfs track new datasets (6f33aef1b1084c2f366ab6fcde05c403511f1ab5)
- add new datasets (e5de966305a3236ebc83263aa73141d98a23209e)
- refactor vep causal eqtl name (fb5e83ea8c8da8784c60c3466547b6947f473443)
- refactor folder name (0e903895e3f050c79d2d7ac3844a2d7e17781219)
- update files to be lfs tracked (98eadaf1a715eed8381c32f3c30e9a5d6899c7f8)
- update main loader script for new datasets (ca412e99e53c186014c69a30e05756f35c4d4024)
- fix file path for causal eqtl data (561af90663da0da6aaee642622e0b661744e742a)
- fix file names for pathogenic datasets (cc38ba3856a08d6401eeff1a36fe5b8dafc1cfe1)
- fix label naming chromatin features (461969927800e16441b1362f60613ed4cb74f12f)
- fix cage file path naming (07eece405aaeb544ff18813d84f9374152352ea8)
- update loading script (1575e22e185371f31365424ed36ba54a547910ff)
- update readme for additional datasets (46d85ff2b9ed6819b26d06fc35d9651d60e51abc)
- update README (66099dcefd0a4c704f334718dafb33c23fefbf74)
- update the return values for cage and regulatory elements (634eabc0c55af9cfa443b4caea908949a20514e3)
- update readme for return values (2f79d04a93ddddeae89b3ad72c9f6c47fc8b5549)
- update the return values for cage (4504d2df698616dd2ed3f3300b7925e6c3036152)
- missing punctuation (ea6154d266f23abce923a09961af9cdc2164e41b)

.gitattributes

@@ -54,3 +54,13 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.jpeg filter=lfs diff=lfs merge=lfs -text
 *.webp filter=lfs diff=lfs merge=lfs -text
 rna_expression_values.csv filter=lfs diff=lfs merge=lfs -text
+chromatin_features/histones_and_dnase_subset.csv filter=lfs diff=lfs merge=lfs -text
+chromatin_features/histones_and_dnase.csv filter=lfs diff=lfs merge=lfs -text
+regulatory_elements/enhancer_dataset.csv filter=lfs diff=lfs merge=lfs -text
+regulatory_elements/enhancer_dataset_subset.csv filter=lfs diff=lfs merge=lfs -text
+regulatory_elements/promoter_dataset.csv filter=lfs diff=lfs merge=lfs -text
+regulatory_elements/promoter_dataset_subset.csv filter=lfs diff=lfs merge=lfs -text
+variant_effect_pathogenic/vep_pathogenic_coding.csv filter=lfs diff=lfs merge=lfs -text
+variant_effect_pathogenic/vep_pathogenic_non_coding.csv filter=lfs diff=lfs merge=lfs -text
+variant_effect_pathogenic/vep_pathogenic_non_coding_subset.csv filter=lfs diff=lfs merge=lfs -text
+variant_effect_causal_eqtl/All_Tissues.csv filter=lfs diff=lfs merge=lfs -text

README.md

@@ -3,30 +3,40 @@ license: cc-by-nc-sa-4.0
 language:
 - en
 tags:
-- biology
-- genomics
-pretty_name: Genomics Long Range Benchmark
+- Genomics
+- Benchmarks
+- Language Models
+- DNA
+pretty_name: Genomics Long-Range Benchmark
 viewer: false
 ---
 
 ## Summary
-The motivation of the genomics long range benchmark (LRB) is to compile a set of biologically relevant genomic tasks requiring long-range dependencies which will act as a robust evaluation tool for genomic language models. 
+The motivation of the genomics long-range benchmark (LRB) is to compile a set of 
+biologically relevant genomic tasks requiring long-range dependencies which will act as a robust evaluation tool for genomic language models. 
 While serving as a strong basis of evaluation, the benchmark must also be efficient and user-friendly. 
 To achieve this we strike a balance between task complexity and computational cost through strategic decisions, such as down-sampling or combining datasets.
 
-## Dataset Tasks
-The Genomics LRB is a collection of tasks which can be loaded by passing in the corresponding `task_name` into the `load_dataset` function. All of the following datasets 
-allow the user to specify an arbitrarily long sequence length, giving more context to the task, by passing `sequence_length` kwarg to `load_dataset`. Additional task specific kwargs, if applicable,
- are mentioend in the sections below.<br>
+## Benchmark Tasks
+The Genomics LRB is a collection of nine tasks which can be loaded by passing in the 
+corresponding `task_name` into the `load_dataset` function. All of the following datasets 
+allow the user to specify an arbitrarily long sequence length, giving more context 
+to the task, by passing the `sequence_length` kwarg to `load_dataset`. Additional task 
+specific kwargs, if applicable, are mentioned in the sections below.<br>
 *Note that as you increase the context length to very large numbers you may start to reduce the size of the dataset since a large context size may 
 cause indexing outside the boundaries of chromosomes.
 
-|        Task         |        `task_name`        |      Sample Output          | # Train Seqs | # Test Seqs |
-|        ---------         |         ----------        |      ------          | ------------ | ----------- |
-| CAGE Prediction | `cage_prediction`|  {sequence, labels, chromosome}  |     36086        |     1922    |
-| Bulk RNA Expression |   `bulk_rna_expression`    |  {sequence, labels, chromosome}  |     22827       |     990     |
-| Variant Effect Gene Expression |   `variant_effect_gene_expression`    |  {ref sequence, alt sequence, label, tissue, chromosome, distance to nearest TSS}  |     89060  |     8862    |
-
+| Task  | `task_name` | Sample Output                                                                             | ML Task Type            | # Outputs   | # Train Seqs | # Test Seqs | Data Source | 
+|-------|-------------|-------------------------------------------------------------------------------------------|-------------------------|-------------|--------------|----------- |----------- |
+| Variant Effect Causal eQTL           | `variant_effect_causal_eqtl`           | {ref sequence, alt sequence, label, tissue, chromosome,position, distance to nearest TSS} | SNP Classification      | 1           | 88717        |     8846    | GTEx (via [Enformer](https://www.nature.com/articles/s41592-021-01252-x)) |
+| Variant Effect Pathogenic ClinVar    | `variant_effect_pathogenic_clinvar`    | {ref sequence, alt sequence, label, chromosome, position}                                 | SNP Classification      | 1           | 38634        |     1018    | ClinVar, gnomAD (via [GPN-MSA](https://www.biorxiv.org/content/10.1101/2023.10.10.561776v1)) |
+| Variant Effect Pathogenic OMIM       | `variant_effect_pathogenic_omim`       | {ref sequence, alt sequence, label,chromosome, position}                                  | SNP Classification      | 1           | -            |     2321473    |OMIM, gnomAD (via [GPN-MSA](https://www.biorxiv.org/content/10.1101/2023.10.10.561776v1))  |
+| CAGE Prediction                      | `cage_prediction`                      | {sequence, labels, chromosome,label_start_position,label_stop_position}                   | Binned Regression       | 50 per bin  | 33891        |     1922    | FANTOM5 (via [Basenji](https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1008050)) |
+| Bulk RNA Expression                  | `bulk_rna_expression`                  | {sequence, labels, chromosome,position}                                                   | Seq-wise Regression     | 218         | 22827        |     990     | GTEx, FANTOM5 (via [ExPecto](https://www.nature.com/articles/s41588-018-0160-6)) |
+| Chromatin Features Histone_Marks     | `chromatin_features_histone_marks`     | {sequence, labels,chromosome, position, label_start_position,label_stop_position}         | Seq-wise Classification | 20          | 2203689      |     227456    | ENCODE, Roadmap Epigenomics (via [DeepSea](https://pubmed.ncbi.nlm.nih.gov/30013180/) |
+| Chromatin Features DNA_Accessibility | `chromatin_features_dna_accessibility` | {sequence, labels,chromosome, position, label_start_position,label_stop_position}         | Seq-wise Classification | 20          | 2203689      | 227456        | ENCODE, Roadmap Epigenomics (via [DeepSea](https://pubmed.ncbi.nlm.nih.gov/30013180/)) |
+| Regulatory Elements Promoter         | `regulatory_element_promoter`          | {sequence, label,chromosome, start, stop, label_start_position,label_stop_position}       | Seq-wise Classification | 1|     953376   |     96240    | SCREEN |
+| Regulatory Elements Enhancer         | `regulatory_element_enhancer`          | {sequence, label,chromosome, start, stop, label_start_position,label_stop_position}       | Seq-wise Classification | 1|     1914575  | 192201      | SCREEN |
 
 ## Usage Example
 ```python
@@ -35,8 +45,13 @@ from datasets import load_dataset
 # Use this parameter to download sequences of arbitrary length (see docs below for edge cases)
 sequence_length=2048
 
-# One of ["cage_prediction", "bulk_rna_expression", "variant_effect_gene_expression"] 
-task_name = "variant_effect_gene_expression"
+# One of:
+# ["variant_effect_causal_eqtl","variant_effect_pathogenic_clinvar",
+# "variant_effect_pathogenic_omim","cage_prediction", "bulk_rna_expression",
+# "chromatin_features_histone_marks","chromatin_features_dna_accessibility",
+# "regulatory_element_promoter","regulatory_element_enhancer"] 
+
+task_name = "variant_effect_causal_eqtl"
 
 dataset = load_dataset(
     "InstaDeepAI/genomics-long-range-benchmark",
@@ -46,20 +61,128 @@ dataset = load_dataset(
 
 ```
 
+### 1. Variant Effect Causal eQTL
+Predicting the effects of genetic variants, particularly expression quantitative trait loci (eQTLs), is essential for understanding the molecular basis of several diseases.
+eQTLs are genomic loci that are associated with variations in mRNA expression levels among individuals.
+By linking genetic variants to causal changes in mRNA expression, researchers can 
+uncover how certain variants contribute to disease development.
+
+#### Source
+Original data comes from GTEx. Processed data in the form of vcf files for positive 
+and negative variants across 49 different tissue types were obtained from the
+[Enformer paper](https://www.nature.com/articles/s41592-021-01252-x) located [here](https://console.cloud.google.com/storage/browser/dm-enformer/data/gtex_fine/vcf?pageState=%28%22StorageObjectListTable%22:%28%22f%22:%22%255B%255D%22%29%29&prefix=&forceOnObjectsSortingFiltering=false). 
+Sequence data originates from the GRCh38 genome assembly.
+
+#### Data Processing
+Fine-mapped GTEx eQTLs originate from [Wang et al](https://www.nature.com/articles/s41467-021-23134-8), while the negative matched set of 
+variants comes from [Avsec et al](https://www.nature.com/articles/s41592-021-01252-x)
+. The statistical fine-mapping tool SuSiE was used to label variants. 
+Variants from the fine-mapped eQTL set were selected and given positive labels if 
+their posterior inclusion probability was > 0.9,
+as assigned by SuSiE. Variants from the matched negative set were given negative labels if their
+posterior inclusion probability was < 0.01.
+
+#### Task Structure
+
+Type: Binary classification<br>
+
+Task Args:<br>
+`sequence_length`: an integer type, the desired final sequence length<br>
+
+Input: a genomic nucleotide sequence centered on the SNP with the reference allele at the SNP location, a genomic nucleotide sequence centered on the SNP with the alternative allele at the SNP location, and tissue type<br>
+Output: a binary value referring to whether the variant has a causal effect on gene 
+expression
+
+#### Splits
+Train: chromosomes 1-8, 11-22, X, Y<br>
+Test: chromosomes 9,10
+
+---
+
+### 2. Variant Effect Pathogenic ClinVar
+A coding variant refers to a genetic alteration that occurs within the protein-coding regions of the genome, also known as exons.
+Such alterations can impact protein structure, function, stability, and interactions 
+with other molecules, ultimately influencing cellular processes and potentially contributing to the development of genetic diseases. 
+Predicting variant pathogenicity is crucial for guiding research into disease mechanisms and personalized treatment strategies, enhancing our ability to understand and manage genetic disorders effectively.
+
+#### Source
+Original data comes from ClinVar and gnomAD. However, we use processed data files 
+from the [GPN-MSA paper](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10592768/) 
+located [here](https://huggingface.co/datasets/songlab/human_variants/blob/main/test.parquet).
+Sequence data originates from the GRCh38 genome assembly.
+
+#### Data Processing
+Positive labels correspond to pathogenic variants originating from ClinVar whose review status was
+described as having at least a single submitted record with a classification but without assertion criteria.
+The negative set are variants that are defined as common from gnomAD. gnomAD version 3.1.2 was downloaded and filtered to variants with allele number of at least 25,000. Common
+variants were defined as those with MAF > 5%.
+
+#### Task Structure
+
+Type: Binary classification<br>
+
+Task Args:<br>
+`sequence_length`: an integer type, the desired final sequence length<br>
+
+Input: a genomic nucleotide sequence centered on the SNP with the reference allele at the SNP location, a genomic nucleotide sequence centered on the SNP with the alternative allele at the SNP location<br>
+Output: a binary value referring to whether the variant is pathogenic or not
+
+#### Splits
+Train: chromosomes 1-7, 9-22, X, Y<br>
+Test: chromosomes 8
+
+---
+
+### 3. Variant Effect Pathogenic OMIM
+Predicting the effects of regulatory variants on pathogenicity is crucial for understanding disease mechanisms.
+Elements that regulate gene expression are often located in non-coding regions, and variants in these areas can disrupt normal cellular function, leading to disease.
+Accurate predictions can identify biomarkers and therapeutic targets, enhancing personalized medicine and genetic risk assessment. 
+
+#### Source
+Original data comes from the Online Mendelian Inheritance in Man (OMIM) and gnomAD 
+databases. 
+However, we use processed data files from the 
+[GPN-MSA paper](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10592768/) located [here](
+https://huggingface.co/datasets/songlab/omim/blob/main/test.parquet).
+Sequence data originates from the GRCh38 genome assembly.
+
+#### Data Processing
+Positive labeled data originates from a curated set of pathogenic variants located 
+in the Online Mendelian Inheritance in Man (OMIM) catalog. The negative set is 
+composed of variants that are defined as common from gnomAD. gnomAD version 3.1.2 was downloaded and filtered to variants with
+allele number of at least 25,000. Common variants were defined as those with minor allele frequency
+(MAF) > 5%. 
 
+#### Task Structure
 
-### 1. CAGE Prediction
-Cap Analysis Gene Expression(CAGE) is a biological assay used to measure the level of mRNA production rather than steady state values, taking into account both production and 
-degradation. Being able to accurately predict mRNA levels as measured by CAGE is essential for deciphering tissue-specific expression patterns, transcriptional networks, and 
-identifying differentially expressed genes with functional significance. 
+Type: Binary classification<br>
+
+Task Args:<br>
+`sequence_length`: an integer type, the desired final sequence length<br>
+`subset`: a boolean type, whether to use the full dataset or a subset of the dataset (we provide this option as the full dataset has millions of samples)
+
+Input: a genomic nucleotide sequence centered on the SNP with the reference allele at the SNP location, a genomic nucleotide sequence centered on the SNP with the alternative allele at the SNP location<br>
+Output: a binary value referring to whether the variant is pathogenic or not
+
+#### Splits
+Test: all chromosomes
+
+---
+
+### 4. CAGE Prediction
+CAGE provides accurate high-throughput measurements of RNA expression by mapping TSSs at a nucleotide-level resolution.
+This is vital for detailed mapping of TSSs, understanding gene regulation mechanisms, and obtaining quantitative expression data to study gene activity comprehensively.
 
 #### Source
-Original CAGE data comes from FANTOM5. We used processed labeled data obtained from the [Basenji paper](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5932613/) which also used to train Enformer and is located [here](https://console.cloud.google.com/storage/browser/basenji_barnyard/data/human?pageState=(%22StorageObjectListTable%22:(%22f%22:%22%255B%255D%22))&prefix=&forceOnObjectsSortingFiltering=false).
+Original CAGE data comes from FANTOM5. We used processed labeled data obtained from 
+the [Basenji paper](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5932613/) which 
+also used to train Enformer and is located [here](https://console.cloud.google.com/storage/browser/basenji_barnyard/data/human?pageState=%28%22StorageObjectListTable%22:%28%22f%22:%22%255B%255D%22%29%29&prefix=&forceOnObjectsSortingFiltering=false).
 Sequence data originates from the GRCh38 genome assembly.
 
 #### Data Processing
 The original dataset from the Basenji paper includes labels for 638 CAGE total tracks over 896 bins (each bin corresponding to 128 base pairs) 
-totaling over ~70 GB. In the interest of dataset size and user friendliness, only a subset of the labels are selected. 
+totaling over ~70 GB. In the interest of dataset size and user-friendliness, only a 
+subset of the labels are selected. 
 From the 638 CAGE tracks, 50 of these tracks are selected with the following criteria:
 
   1. Only select one cell line
@@ -67,6 +190,7 @@ From the 638 CAGE tracks, 50 of these tracks are selected with the following cri
   3. Only select one donor
      
 The [896 bins, 50 tracks] labels total in at ~7 GB. A description of the 50 included CAGE tracks can be found here `cage_prediction/label_mapping.csv`.
+*Note the data in this repository for this task has not already been log(1+x) normalized. 
 
 #### Task Structure
 
@@ -75,7 +199,7 @@ Because this task involves predicting expression levels for 128bp bins and there
 you request a sequence length smaller than 114,688 bps than the labels will be subsetted.
 
 Task Args:<br>
-`sequence_length`: an interger type, the desired final sequence length, *must be a multiple of 128 given the binned nature of labels<br>
+`sequence_length`: an integer type, the desired final sequence length, *must be a multiple of 128 given the binned nature of labels<br>
 
 Input: a genomic nucleotide sequence<br>
 Output: a variable length vector depending on the requested sequence length [requested_sequence_length / 128, 50]
@@ -84,30 +208,36 @@ Output: a variable length vector depending on the requested sequence length [req
 Train/Test splits were maintained from Basenji and Enformer where randomly sampling was used to generate the splits. Note that for this dataset a validation set is also returned. In practice we merged the validation 
 set with the train set and use cross validation to select a new train and validation set from this combined set.
 
-#### Metrics
-Mean Pearson correlation across tracks - compute Pearson correlation for a track using all positions for all genes in the test set, then mean over all tracks <br>
-Mean Pearson correlation across genes - compute Pearson correlation for a gene using all positions and all tracks, then mean over all genes in the test set <br>
-R<sup>2</sup>
-
 
 ---
 
-### 2. Bulk RNA Expression
-In comparison to CAGE, bulk RNA sequencing assays measure the steady state level (both transription and degradation) of mRNA in a population of cells.
+### 5. Bulk RNA Expression
+Gene expression involves the process by which information encoded in a gene directs the synthesis of a functional gene product, typically a protein, through transcription and translation.
+Transcriptional regulation determines the amount of mRNA produced, which is then translated into proteins. Developing a model that can predict RNA expression levels solely from sequence 
+data is crucial for advancing our understanding of gene regulation, elucidating disease mechanisms, and identifying functional sequence variants.
 
 #### Source
 Original data comes from GTEx. We use processed data files from the [ExPecto paper](https://www.nature.com/articles/s41588-018-0160-6) found 
 [here](https://github.com/FunctionLab/ExPecto/tree/master/resources). Sequence data originates from the GRCh37/hg19 genome assembly.
 
 #### Data Processing
-The continuous labels were log(1+x) transformed and standardized. A list of names of tissues corresponding to the labels can be found here: `bulk_rna_expression/label_mapping.csv`.
+The authors of ExPecto determined representative TSS for Pol II transcribed genes 
+based on quantification of CAGE reads from the FANTOM5 project. The specific procedure they used is as
+follows, a CAGE peak was associated to a GENCODE gene if it was withing 1000 bps from a
+GENCODE v24 annotated TSS. The most abundant CAGE peak for each gene was then selected
+as the representative TSS. When no CAGE peak could be assigned to a gene, the annotated gene
+start position was used as the representative TSS. We log(1 + x) normalized then standardized the
+RNA-seq counts before training models. A list of names of tissues corresponding to 
+the labels can be found here: `bulk_rna_expression/label_mapping.csv`. *Note the 
+data in this repository for this task has already been log(1+x) normalized and 
+standardized to mean 0 and unit variance.
 
 #### Task Structure
 
 Type: Multi-variable regression<br>
 
 Task Args:<br>
-`sequence_length`: an interger type, the desired final sequence length<br>
+`sequence_length`: an integer type, the desired final sequence length<br>
 
 Input: a genomic nucleotide sequence centered around the CAGE representative trancription start site<br>
 Output: a 218 length vector of continuous values corresponding to the bulk RNA expression levels in 218 different tissue types
@@ -116,42 +246,70 @@ Output: a 218 length vector of continuous values corresponding to the bulk RNA e
 Train: chromosomes 1-7,9-22,X,Y<br>
 Test: chromosome 8
 
-#### Metrics
-Mean Spearman correlation across tissues <br>
-Mean Spearman correlation across genes <br>
-R<sup>2</sup>
-
 ---
-
-### 3. Variant Effect Gene Expression
-In genomics, a key objective is to predict how genetic variants affect gene expression in specific cell types.
+### 6. Chromatin Features 
+Predicting chromatin features, such as histone marks and DNA accessibility, is crucial for understanding gene regulation, as these features indicate chromatin state and are essential for transcription activation.
 
 #### Source
-Original data comes from GTEx. However, we used processed data files from the [Enformer paper](https://www.nature.com/articles/s41592-021-01252-x) located [here](https://console.cloud.google.com/storage/browser/dm-enformer/data/gtex_fine/vcf?pageState=(%22StorageObjectListTable%22:(%22f%22:%22%255B%255D%22))&prefix=&forceOnObjectsSortingFiltering=false). 
-Sequence data originates from the GRCh38 genome assembly.
+Original data used to generate labels for histone marks and DNase profiles comes from the ENCODE and Roadmap Epigenomics project. We used processed data files from the [Deep Sea paper](https://www.nature.com/articles/nmeth.3547) to build this dataset.
+Sequence data originates from the GRCh37/hg19 genome assembly.
 
 #### Data Processing
-In Enformer the datasets were partitioned in 48 different sets based on the tissue types. In our framing of the task we combine all samples across all tissues into one set 
-and provide the tissue type along with each sample.
-
-As data files were used from Enformer, the labels were constructed according to their methodology - variants were labeled as 1 if their posterior inclusion probability was greater than 0.9 as 
-assigned by the population-based fine-mapping tool SuSiE, while a matched set of negative variants was built with posterior inclusion probabilities of less than .01.
+The authors of DeepSea processed the data by chunking the human genome
+into 200 bp bins where for each bin labels were determined for hundreds of different chromatin
+features. Only bins with at least one transcription factor binding event were 
+considered for the dataset. If the bin overlapped with a peak region of the specific 
+chromatin profile by more than half of the
+sequence, a positive label was assigned. DNA sequences were obtained from the human reference
+genome assembly GRCh37. To make the dataset more accessible, we randomly sub-sampled the
+chromatin profiles from 125 to 20 tracks for the histones dataset and from 104 to 20 tracks for the
+DNA accessibility dataset.
 
 #### Task Structure
 
-Type: Binary classification<br>
+Type: Multi-label binary classification
 
 Task Args:<br>
-`sequence_length`: an interger type, the desired final sequence length<br>
+`sequence_length`: an integer type, the desired final sequence length<br>
+`subset`: a boolean type, whether to use the full dataset or a subset of the dataset (we provide this option as the full dataset has millions of samples)
 
-Input: a genomic nucleotide sequence centered on the SNP with the reference allele at the SNP location, a genomic nucleotide sequence centered on the SNP with the alternative allele at the SNP location, and tissue type<br>
-Output: a binary value refering to whether the variant has an effect on gene expression 
+Input: a genomic nucleotide sequence centered on the 200 base pair bin that is associated with the labels<br>
+Output: a vector of length 20 with binary entries
 
 #### Splits
-Train: chromosomes 1-8, 11-22, X, Y<br>
-Test: chromosomes 9,10
+Train set: chromosomes 1-7,10-22<br>
+Test set: chromosomes 8,9
+
+---
+### 7. Regulatory Elements
+Cis-regulatory elements, such as promoters and enhancers, control the spatial and temporal expression of genes. 
+These elements are essential for understanding gene regulation mechanisms and how genetic variations can lead to differences in gene expression.
+
+#### Source
+Original data annotations to build labels came from the Search Candidate cis-Regulatory Elements by ENCODE project. Sequence data originates from the GRCh38 
+genome assembly.
 
-#### Metrics
-Accuracy<br>
-AUROC<br>
-AUPRC
+#### Data Processing
+The data is processed as follows, we break the human
+reference genome into 200 bp non-overlapping chunks. If the 200 bp chunk overlaps by at least 50%
+or more with a contiguous region from the set of annotated cis-regulatory elements (promoters or
+enhancers), we label them as positive, else the chunk is labeled as negative. The resulting dataset
+was composed of ∼15M negative samples and ∼50k positive promoter samples and ∼1M positive
+enhancer samples. We randomly sub-sampled the negative set to 1M samples, and kept 
+all positive
+samples, to make this dataset more manageable in size.
+
+#### Task Structure
+
+Type: Binary classification
+
+Task Args:<br>
+`sequence_length`: an integer type, the desired final sequence length<br>
+`subset`: a boolean type, whether to use the full dataset or a subset of the dataset (we provide this option as the full dataset has millions of samples)
+
+Input: a genomic nucleotide sequence centered on the 200 base pair bin that is associated with the label<br>
+Output: a single binary value
+
+#### Splits
+Train set: chromosomes 1-7,10-22<br>
+Test set: chromosomes 8,9

chromatin_features/histones_and_dnase.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d6cc04691aca70c876018f15463ba697ddd790af8acda7bbdf14417a3032d153
+size 356382794

chromatin_features/histones_and_dnase_subset.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e5a63d4067cd0f1da65aecb6b1ed0a2e1beb19a104f61bfc06ca401b0c14dc14
+size 47732536

genomics-long-range-benchmark.py

@@ -14,14 +14,12 @@ import pandas as pd
 from datasets import DatasetInfo
 from pyfaidx import Fasta
 from abc import ABC, abstractmethod
-from Bio.Seq import Seq
-from Bio import SeqIO
-import pysam
+
 
 """
---------------------------------------------------------------------------------------------
+----------------------------------------------------------------------------------------
 Reference Genome URLS:
--------------------------------------------------------------------------------------------
+----------------------------------------------------------------------------------------
 """
 H38_REFERENCE_GENOME_URL = (
     "https://hgdownload.soe.ucsc.edu/goldenPath/hg38/bigZips/" "hg38.fa.gz"
@@ -31,9 +29,9 @@ H19_REFERENCE_GENOME_URL = (
 )
 
 """
---------------------------------------------------------------------------------------------
+----------------------------------------------------------------------------------------
 Task Specific Handlers:
--------------------------------------------------------------------------------------------
+----------------------------------------------------------------------------------------
 """
 
 class GenomicLRATaskHandler(ABC):
@@ -97,8 +95,8 @@ class GenomicLRATaskHandler(ABC):
 
     def download_and_extract_gz(self, file_url, cache_dir_root):
         """
-        Downloads and extracts a gz file into the given cache directory. Returns the full file path
-        of the extracted gz file.
+        Downloads and extracts a gz file into the given cache directory. Returns the
+        full file path of the extracted gz file.
         Args:
             file_url: url of the gz file to be downloaded and extracted.
             cache_dir_root: Directory to extract file into.
@@ -138,29 +136,30 @@ class CagePredictionHandler(GenomicLRATaskHandler):
         50,
     )  # 50 is a subset of CAGE tracks from the original enformer dataset
     NPZ_SPLIT = 1000  # number of files per npz file.
-    NUM_BP_PER_BIN = 128 # number of base pairs per bin in labels
+    NUM_BP_PER_BIN = 128  # number of base pairs per bin in labels
 
     def __init__(self, sequence_length=DEFAULT_LENGTH, **kwargs):
         """
         Creates a new handler for the CAGE task.
         Args:
-            sequence_length: allows for increasing sequence context. Sequence length must be a multiple of 128 to align with binned labels. Note: increasing
-            sequence length may decrease the number of usable samples.
+            sequence_length: allows for increasing sequence context. Sequence length
+            must be an even multiple of 128 to align with binned labels. Note:
+            increasing sequence length may decrease the number of usable samples.
         """
         self.reference_genome = None
         self.coordinate_csv_file = None
         self.target_files_by_split = {}
 
+
         assert (sequence_length // 128) % 2 == 0, (
-            f"Requested sequence length must be an even multuple of 128 to align with the binned labels."
+            f"Requested sequence length must be an even multuple of 128 to align "
+            f"with the binned labels."
         )
 
         self.sequence_length = sequence_length
 
         if self.sequence_length < self.DEFAULT_LENGTH:
-
-            self.TARGET_SHAPE = (self.sequence_length//128,50)
-    
+            self.TARGET_SHAPE = (self.sequence_length // 128, 50)
 
     def get_info(self, description: str) -> DatasetInfo:
         """
@@ -174,7 +173,11 @@ class CagePredictionHandler(GenomicLRATaskHandler):
                 # array of sequence length x num_labels
                 "labels": datasets.Array2D(shape=self.TARGET_SHAPE, dtype="float32"),
                 # chromosome number
-                "chromosome":datasets.Value(dtype="string")
+                "chromosome": datasets.Value(dtype="string"),
+                # start
+                "labels_start": datasets.Value(dtype="int32"),
+                # stop
+                "labels_stop": datasets.Value(dtype="int32")
             }
         )
         return datasets.DatasetInfo(
@@ -192,7 +195,7 @@ class CagePredictionHandler(GenomicLRATaskHandler):
         """
 
         # Manually download the reference genome since there are difficulties when
-        # streaming the dataset
+        # streaming
         reference_genome_file = self.download_and_extract_gz(
             H38_REFERENCE_GENOME_URL, cache_dir_root
         )
@@ -225,7 +228,6 @@ class CagePredictionHandler(GenomicLRATaskHandler):
         self.target_files_by_split["test"] = test_file_dict
         self.target_files_by_split["validation"] = valid_file_dict
 
-
         return [
             datasets.SplitGenerator(
                 name=datasets.Split.TRAIN,
@@ -241,7 +243,6 @@ class CagePredictionHandler(GenomicLRATaskHandler):
             ),
         ]
 
-
     def generate_examples(self, split):
         """
         A generator which produces examples for the given split, each with a sequence
@@ -250,24 +251,28 @@ class CagePredictionHandler(GenomicLRATaskHandler):
         """
 
         target_files = self.target_files_by_split[split]
-        sequence_length = self.sequence_length
 
         key = 0
         coordinates_dataframe = pd.read_csv(self.coordinate_csv_file)
         filtered = coordinates_dataframe[coordinates_dataframe["split"] == split]
         for sequential_idx, row in filtered.iterrows():
             start, stop = int(row["start"]) - 1, int(
-                row["stop"]) - 1  # -1 since vcf coords are 1-based
+                row["stop"]) - 1  # -1 since coords are 1-based
 
             chromosome = row['chrom']
-            
-            padded_sequence = pad_sequence(
+
+            padded_sequence,new_start,new_stop = pad_sequence(
                 chromosome=self.reference_genome[chromosome],
                 start=start,
-                sequence_length=sequence_length,
+                sequence_length=self.sequence_length,
                 end=stop,
+                return_new_start_stop=True
             )
 
+            if self.sequence_length >= self.DEFAULT_LENGTH:
+                new_start = start
+                new_stop = stop
+
             # floor npy_idx to the nearest 1000
             npz_file = np.load(
                 target_files[int((row["npy_idx"] // self.NPZ_SPLIT) * self.NPZ_SPLIT)]
@@ -277,21 +282,22 @@ class CagePredictionHandler(GenomicLRATaskHandler):
                     split == "validation"
             ):  # npy files are keyed by ["train", "test", "valid"]
                 split = "valid"
-            targets = npz_file[f"target-{split}-{row['npy_idx']}.npy"][0] # select 0 since there is extra dimension
-            
-            
+            targets = npz_file[f"target-{split}-{row['npy_idx']}.npy"][
+                0]  # select 0 since there is extra dimension
+
             # subset the targets if sequence length is smaller than 114688 (
             # DEFAULT_LENGTH)
             if self.sequence_length < self.DEFAULT_LENGTH:
                 idx_diff = (self.DEFAULT_LENGTH - self.sequence_length) // 2 // 128
                 targets = targets[idx_diff:-idx_diff]
 
-            
             if padded_sequence:
                 yield key, {
                     "labels": targets,
                     "sequence": standardize_sequence(padded_sequence),
-                    "chromosome": re.sub("chr","",chromosome)
+                    "chromosome": re.sub("chr", "", chromosome),
+                    "labels_start": new_start,
+                    "labels_stop": new_stop
                 }
                 key += 1
 
@@ -325,7 +331,7 @@ class BulkRnaExpressionHandler(GenomicLRATaskHandler):
     Handler for the Bulk RNA Expression task.
     """
 
-    DEFAULT_LENGTH = 114688
+    DEFAULT_LENGTH = 100000
 
     def __init__(self, sequence_length=DEFAULT_LENGTH, **kwargs):
         """
@@ -351,7 +357,9 @@ class BulkRnaExpressionHandler(GenomicLRATaskHandler):
                 # list of expression values in each tissue
                 "labels": datasets.Sequence(datasets.Value("float32")),
                 # chromosome number
-                "chromosome":datasets.Value(dtype="string")
+                "chromosome": datasets.Value(dtype="string"),
+                # position
+                "position": datasets.Value(dtype="int32"),
             }
         )
         return datasets.DatasetInfo(
@@ -368,7 +376,7 @@ class BulkRnaExpressionHandler(GenomicLRATaskHandler):
         The Bulk RNA Expression dataset requires the reference hg19 genome, coordinate
         csv file,and label csv file to be saved.
         """
-        # Manually download the reference genome since there are difficulties when streaming
+
         reference_genome_file = self.download_and_extract_gz(
             H19_REFERENCE_GENOME_URL, cache_dir_root
         )
@@ -398,7 +406,7 @@ class BulkRnaExpressionHandler(GenomicLRATaskHandler):
         key = 0
         for idx, coordinates_row in coordinates_split_df.iterrows():
             start = coordinates_row[
-                        "CAGE_representative_TSS"] - 1  # -1 since vcf coords are 1-based
+                        "CAGE_representative_TSS"] - 1  # -1 since coords are 1-based
 
             chromosome = coordinates_row["chrom"]
             labels_row = labels_df.loc[idx].values
@@ -412,21 +420,22 @@ class BulkRnaExpressionHandler(GenomicLRATaskHandler):
                 yield key, {
                     "labels": labels_row,
                     "sequence": standardize_sequence(padded_sequence),
-                    "chromosome":re.sub("chr","",chromosome)
+                    "chromosome": re.sub("chr", "", chromosome),
+                    "position": coordinates_row["CAGE_representative_TSS"]
                 }
                 key += 1
 
 
-class VariantEffectPredictionHandler(GenomicLRATaskHandler):
+class VariantEffectCausalEqtl(GenomicLRATaskHandler):
     """
-    Handler for the Variant Effect Prediction task.
+    Handler for the Variant Effect Causal eQTL task.
     """
 
-    DEFAULT_LENGTH = 114688
+    DEFAULT_LENGTH = 100000
 
     def __init__(self, sequence_length=DEFAULT_LENGTH, **kwargs):
         """
-        Creates a new handler for the Variant Effect Prediction Task.
+        Creates a new handler for the Variant Effect Causal eQTL Task.
         Args:
             sequence_length: Length of the sequence to pad around the SNP position
 
@@ -436,9 +445,9 @@ class VariantEffectPredictionHandler(GenomicLRATaskHandler):
 
     def get_info(self, description: str) -> DatasetInfo:
         """
-        Returns the DatasetInfo for the Variant Effect Prediction dataset. Each example
-        includes a  genomic sequence with the reference allele as well as the genomic sequence with the alternative allele,
-        and a binary label.
+        Returns the DatasetInfo for the Variant Effect Causal eQTL dataset. Each example
+        includes a  genomic sequence with the reference allele as well as the genomic
+        sequence with the alternative allele, and a binary label.
         """
         features = datasets.Features(
             {
@@ -451,8 +460,10 @@ class VariantEffectPredictionHandler(GenomicLRATaskHandler):
                 "tissue": datasets.Value(dtype="string"),
                 # chromosome number
                 "chromosome": datasets.Value(dtype="string"),
+                # variant position
+                "position": datasets.Value(dtype="int32"),
                 # distance to nearest tss
-                "distance_to_nearest_tss":datasets.Value(dtype="int32")
+                "distance_to_nearest_tss": datasets.Value(dtype="int32")
             }
         )
 
@@ -478,7 +489,7 @@ class VariantEffectPredictionHandler(GenomicLRATaskHandler):
 
         self.reference_genome = Fasta(reference_genome_file, one_based_attributes=False)
         self.coordinates_labels_csv_file = dl_manager.download_and_extract(
-            f"variant_effect_gene_expression/All_Tissues.csv"
+            f"variant_effect_causal_eqtl/All_Tissues.csv"
         )
 
         return super().split_generators(dl_manager, cache_dir_root)
@@ -496,7 +507,7 @@ class VariantEffectPredictionHandler(GenomicLRATaskHandler):
 
         key = 0
         for idx, row in coordinates_split_df.iterrows():
-            start = row["POS"] - 1  # sub 1 to create idx since vcf coords are 1-based
+            start = row["POS"] - 1  # sub 1 to create idx since coords are 1-based
             alt_allele = row["ALT"]
             label = row["label"]
             tissue = row['tissue']
@@ -513,8 +524,8 @@ class VariantEffectPredictionHandler(GenomicLRATaskHandler):
 
             # only if a valid sequence returned
             if ref_forward:
-                # Mutate sequence with the alt allele at the SNP position, which is always
-                # centered in the string returned from pad_sequence
+                # Mutate sequence with the alt allele at the SNP position,
+                # which is always centered in the string returned from pad_sequence
                 alt_forward = list(ref_forward)
                 alt_forward[self.sequence_length // 2] = alt_allele
                 alt_forward = "".join(alt_forward)
@@ -525,14 +536,354 @@ class VariantEffectPredictionHandler(GenomicLRATaskHandler):
                     "chromosome": re.sub("chr", "", chromosome),
                     "ref_forward_sequence": standardize_sequence(ref_forward),
                     "alt_forward_sequence": standardize_sequence(alt_forward),
-                    "distance_to_nearest_tss": distance
+                    "distance_to_nearest_tss": distance,
+                    "position": row["POS"]
                 }
                 key += 1
 
+
+class VariantEffectPathogenicHandler(GenomicLRATaskHandler):
+    """
+    Handler for the Variant Effect Pathogenic Prediction tasks.
+    """
+
+    DEFAULT_LENGTH = 100000
+
+    def __init__(self, sequence_length=DEFAULT_LENGTH, task_name=None, subset=False,
+                 **kwargs):
+        """
+        Creates a new handler for the Variant Effect Pathogenic Tasks.
+        Args:
+            sequence_length: Length of the sequence to pad around the SNP position
+            subset: Whether to return a pre-determined subset of the data.
+
+        """
+        self.sequence_length = sequence_length
+
+        if task_name == 'variant_effect_pathogenic_clinvar':
+            self.data_file_name = "variant_effect_pathogenic/vep_pathogenic_coding.csv"
+        elif task_name == 'variant_effect_pathogenic_omim':
+            self.data_file_name = "variant_effect_pathogenic/" \
+                                  "vep_pathogenic_non_coding_subset.csv" \
+                if subset else "variant_effect_pathogenic/vep_pathogenic_non_coding.csv"
+
+    def get_info(self, description: str) -> DatasetInfo:
+        """
+        Returns the DatasetInfo for the Variant Effect Pathogenic datasets. Each example
+        includes a  genomic sequence with the reference allele as well as the genomic
+        sequence with the alternative allele, and a binary label.
+        """
+        features = datasets.Features(
+            {
+                # DNA sequence
+                "ref_forward_sequence": datasets.Value("string"),
+                "alt_forward_sequence": datasets.Value("string"),
+                # binary label
+                "label": datasets.Value(dtype="int8"),
+                # chromosome number
+                "chromosome": datasets.Value(dtype="string"),
+                # position
+                "position": datasets.Value(dtype="int32")
+            }
+        )
+
+        return datasets.DatasetInfo(
+            # This is the description that will appear on the datasets page.
+            description=description,
+            # This defines the different columns of the dataset and their types
+            features=features,
+        )
+
+    def split_generators(self, dl_manager, cache_dir_root):
+        """
+        Separates files by split and stores filenames in instance variables.
+        The variant effect prediction datasets require the reference hg38 genome and
+        coordinates_labels_csv_file to be saved.
+        """
+
+        reference_genome_file = self.download_and_extract_gz(
+            H38_REFERENCE_GENOME_URL, cache_dir_root
+        )
+
+        self.reference_genome = Fasta(reference_genome_file, one_based_attributes=False)
+        self.coordinates_labels_csv_file = dl_manager.download_and_extract(
+            self.data_file_name)
+
+        if 'non_coding' in self.data_file_name:
+            return [
+                datasets.SplitGenerator(
+                    name=datasets.Split.TEST,
+                    gen_kwargs={"handler": self, "split": "test"}
+                ), ]
+        else:
+            return super().split_generators(dl_manager, cache_dir_root)
+
+    def generate_examples(self, split):
+        """
+        A generator which produces examples each with ref/alt allele
+        and corresponding binary label. The sequences are extended to
+        the desired sequence length and standardized before returning.
+        """
+
+        coordinates_df = pd.read_csv(self.coordinates_labels_csv_file)
+        coordinates_split_df = coordinates_df[coordinates_df["split"] == split]
+
+        key = 0
+        for idx, row in coordinates_split_df.iterrows():
+            start = row["POS"] - 1  # sub 1 to create idx since coords are 1-based
+            alt_allele = row["ALT"]
+            label = row["INT_LABEL"]
+            chromosome = row["CHROM"]
+
+            # get reference forward sequence
+            ref_forward = pad_sequence(
+                chromosome=self.reference_genome[chromosome],
+                start=start,
+                sequence_length=self.sequence_length,
+                negative_strand=False,
+            )
+
+            # only if a valid sequence returned
+            if ref_forward:
+                # Mutate sequence with the alt allele at the SNP position,
+                # which is always centered in the string returned from pad_sequence
+                alt_forward = list(ref_forward)
+                alt_forward[self.sequence_length // 2] = alt_allele
+                alt_forward = "".join(alt_forward)
+
+                yield key, {
+                    "label": label,
+                    "chromosome": re.sub("chr", "", chromosome),
+                    "ref_forward_sequence": standardize_sequence(ref_forward),
+                    "alt_forward_sequence": standardize_sequence(alt_forward),
+                    "position": row['POS']
+                }
+                key += 1
+
+
+class ChromatinFeaturesHandler(GenomicLRATaskHandler):
+    """
+    Handler for the histone marks and DNA accessibility tasks also referred to
+    collectively as Chromatin features.
+    """
+
+    DEFAULT_LENGTH = 100000
+
+    def __init__(self, task_name=None, sequence_length=DEFAULT_LENGTH, subset=False,
+                 **kwargs):
+        """
+        Creates a new handler for the Deep Sea Histone and DNase tasks.
+        Args:
+            sequence_length: Length of the sequence around and including the
+            annotated 200bp bin
+            subset: Whether to return a pre-determined subset of the entire dataset.
+
+        """
+        self.sequence_length = sequence_length
+
+        if sequence_length < 200:
+            raise ValueError(
+                'Sequence length for this task must be greater or equal to 200 bp')
+
+        if 'histone' in task_name:
+            self.label_name = 'HISTONES'
+        elif 'dna' in task_name:
+            self.label_name = 'DNASE'
+
+        self.data_file_name = "chromatin_features/histones_and_dnase_subset.csv" if \
+            subset else "chromatin_features/histones_and_dnase.csv"
+
+    def get_info(self, description: str) -> DatasetInfo:
+        """
+        Returns the DatasetInfo for the histone marks and dna accessibility datasets.
+        Each example includes a genomic sequence and a list of label values.
+        """
+        features = datasets.Features(
+            {
+                # DNA sequence
+                "sequence": datasets.Value("string"),
+                # list of binary chromatin marks
+                "labels": datasets.Sequence(datasets.Value("int8")),
+                # chromosome number
+                "chromosome": datasets.Value(dtype="string"),
+                # starting position in genome which corresponds to label
+                "label_start": datasets.Value(dtype="int32"),
+                # end position in genome which corresponds to label
+                "label_stop": datasets.Value(dtype="int32"),
+            }
+        )
+        return datasets.DatasetInfo(
+            # This is the description that will appear on the datasets page.
+            description=description,
+            # This defines the different columns of the dataset and their types
+            features=features,
+
+        )
+
+    def split_generators(self, dl_manager, cache_dir_root):
+        """
+        Separates files by split and stores filenames in instance variables.
+        The histone marks and dna accessibility datasets require the reference hg19
+        genome and coordinate csv file to be saved.
+        """
+        reference_genome_file = self.download_and_extract_gz(
+            H19_REFERENCE_GENOME_URL, cache_dir_root
+        )
+        self.reference_genome = Fasta(reference_genome_file, one_based_attributes=False)
+
+        self.coordinate_csv_file = dl_manager.download_and_extract(self.data_file_name)
+
+        return super().split_generators(dl_manager, cache_dir_root)
+
+    def generate_examples(self, split):
+        """
+        A generator which produces examples for the given split, each with a sequence
+        and the corresponding labels. The sequences are padded to the correct sequence
+        length and standardized before returning.
+        """
+        coordinates_df = pd.read_csv(self.coordinate_csv_file)
+        coordinates_split_df = coordinates_df[coordinates_df["split"] == split]
+
+        key = 0
+        for idx, coordinates_row in coordinates_split_df.iterrows():
+            start = coordinates_row['POS'] - 1  # -1 since saved coords are 1-based
+            chromosome = coordinates_row["CHROM"]
+
+            # literal eval used since lists are saved as strings in csv
+            labels_row = literal_eval(coordinates_row[self.label_name])
+
+            padded_sequence = pad_sequence(
+                chromosome=self.reference_genome[chromosome],
+                start=start,
+                sequence_length=self.sequence_length,
+            )
+            if padded_sequence:
+                yield key, {
+                    "labels": labels_row,
+                    "sequence": standardize_sequence(padded_sequence),
+                    "chromosome": re.sub("chr", "", chromosome),
+                    "label_start": coordinates_row['POS']-100,
+                    "label_stop": coordinates_row['POS'] + 99,
+                }
+                key += 1
+
+
+class RegulatoryElementHandler(GenomicLRATaskHandler):
+    """
+    Handler for the Regulatory Element Prediction tasks.
+    """
+    DEFAULT_LENGTH = 100000
+
+    def __init__(self, task_name=None, sequence_length=DEFAULT_LENGTH, subset=False,
+                 **kwargs):
+        """
+        Creates a new handler for the Regulatory Element Prediction tasks.
+        Args:
+            sequence_length: Length of the sequence around the element/non-element
+            subset: Whether to return a pre-determined subset of the entire dataset.
+
+        """
+
+        if sequence_length < 200:
+            raise ValueError(
+                'Sequence length for this task must be greater or equal to 200 bp')
+
+        self.sequence_length = sequence_length
+
+        if 'promoter' in task_name:
+            self.data_file_name = 'regulatory_elements/promoter_dataset'
+
+        elif 'enhancer' in task_name:
+            self.data_file_name = 'regulatory_elements/enhancer_dataset'
+
+        if subset:
+            self.data_file_name += '_subset.csv'
+        else:
+            self.data_file_name += '.csv'
+
+    def get_info(self, description: str) -> DatasetInfo:
+        """
+        Returns the DatasetInfo for the Regulatory Element Prediction Tasks.
+        Each example includes a genomic sequence and a label.
+        """
+        features = datasets.Features(
+            {
+                # DNA sequence
+                "sequence": datasets.Value("string"),
+                # label corresponding to whether the sequence has
+                # the regulatory element of interest or not
+                "labels": datasets.Value("int8"),
+                # chromosome number
+                "chromosome": datasets.Value(dtype="string"),
+                # start
+                "label_start": datasets.Value(dtype="int32"),
+                # stop
+                "label_stop": datasets.Value(dtype="int32"),
+            }
+        )
+        return datasets.DatasetInfo(
+            # This is the description that will appear on the datasets page.
+            description=description,
+            # This defines the different columns of the dataset and their types
+            features=features,
+
+        )
+
+    def split_generators(self, dl_manager, cache_dir_root):
+        """
+        Separates files by split and stores filenames in instance variables.
+        """
+        reference_genome_file = self.download_and_extract_gz(
+            H38_REFERENCE_GENOME_URL, cache_dir_root
+        )
+        self.reference_genome = Fasta(reference_genome_file, one_based_attributes=False)
+
+        self.coordinate_csv_file = dl_manager.download_and_extract(
+            self.data_file_name
+        )
+
+        return super().split_generators(dl_manager, cache_dir_root)
+
+    def generate_examples(self, split):
+        """
+        A generator which produces examples for the given split, each with a sequence
+        and the corresponding label. The sequences are padded to the correct sequence
+        length and standardized before returning.
+        """
+        coordinates_df = pd.read_csv(self.coordinate_csv_file)
+
+        coordinates_split_df = coordinates_df[coordinates_df["split"] == split]
+
+        key = 0
+        for _, coordinates_row in coordinates_split_df.iterrows():
+            start = coordinates_row["START"] - 1  # -1 since vcf coords are 1-based
+            end = coordinates_row["STOP"] - 1  # -1 since vcf coords are 1-based
+            chromosome = coordinates_row["CHROM"]
+
+            label = coordinates_row['label']
+
+            padded_sequence = pad_sequence(
+                chromosome=self.reference_genome[chromosome],
+                start=start,
+                end=end,
+                sequence_length=self.sequence_length,
+            )
+
+            if padded_sequence:
+                yield key, {
+                    "labels": label,
+                    "sequence": standardize_sequence(padded_sequence),
+                    "chromosome": re.sub("chr", "", chromosome),
+                    "label_start": coordinates_row["START"],
+                    "label_stop": coordinates_row["STOP"]
+                }
+                key += 1
+
+
 """
---------------------------------------------------------------------------------------------
+----------------------------------------------------------------------------------------
 Dataset loader:
--------------------------------------------------------------------------------------------
+----------------------------------------------------------------------------------------
 """
 
 _DESCRIPTION = """
@@ -542,7 +893,13 @@ Dataset for benchmark of genomic deep learning models.
 _TASK_HANDLERS = {
     "cage_prediction": CagePredictionHandler,
     "bulk_rna_expression": BulkRnaExpressionHandler,
-    "variant_effect_gene_expression": VariantEffectPredictionHandler,
+    "variant_effect_causal_eqtl": VariantEffectCausalEqtl,
+    "variant_effect_pathogenic_clinvar": VariantEffectPathogenicHandler,
+    "variant_effect_pathogenic_omim": VariantEffectPathogenicHandler,
+    "chromatin_features_histone_marks": ChromatinFeaturesHandler,
+    "chromatin_features_dna_accessibility": ChromatinFeaturesHandler,
+    "regulatory_element_promoter": RegulatoryElementHandler,
+    "regulatory_element_enhancer": RegulatoryElementHandler,
 }
 
 
@@ -558,7 +915,7 @@ class GenomicsLRAConfig(datasets.BuilderConfig):
             **kwargs: keyword arguments forwarded to super.
         """
         super().__init__()
-        self.handler = _TASK_HANDLERS[task_name](task_name=task_name,**kwargs)
+        self.handler = _TASK_HANDLERS[task_name](task_name=task_name, **kwargs)
 
 
 # DatasetBuilder
@@ -592,9 +949,9 @@ class GenomicsLRATasks(datasets.GeneratorBasedBuilder):
 
 
 """
---------------------------------------------------------------------------------------------
+----------------------------------------------------------------------------------------
 Global Utils:
--------------------------------------------------------------------------------------------
+----------------------------------------------------------------------------------------
 """
 
 
@@ -613,7 +970,8 @@ def standardize_sequence(sequence: str):
     return sequence
 
 
-def pad_sequence(chromosome, start, sequence_length, end=None, negative_strand=False):
+def pad_sequence(chromosome, start, sequence_length, end=None, negative_strand=False,
+                 return_new_start_stop=False):
     """
     Extends a given sequence to length sequence_length. If
     padding to the given length is outside the gene, returns
@@ -625,7 +983,8 @@ def pad_sequence(chromosome, start, sequence_length, end=None, negative_strand=F
             remainder is added to the end of the sequence.
         end: End index of original sequence. If no end is specified, it creates a
             centered sequence around the start index.
-        negative_strand: If negative_strand, returns the reverse compliment of the sequence
+        negative_strand: If negative_strand, returns the reverse compliment of the
+        sequence
     """
     if end:
         pad = (sequence_length - (end - start)) // 2
@@ -639,5 +998,12 @@ def pad_sequence(chromosome, start, sequence_length, end=None, negative_strand=F
     if start < 0 or end >= len(chromosome):
         return
     if negative_strand:
+        if return_new_start_stop:
+            return chromosome[start:end].reverse.complement.seq ,start, end
+
         return chromosome[start:end].reverse.complement.seq
+
+    if return_new_start_stop:
+        return chromosome[start:end].seq , start, end
+
     return chromosome[start:end].seq

regulatory_elements/enhancer_dataset.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a511212c76aba8c25ecd372531fe78767cbd4065ce72d4899dcb4ba1429750c
+size 66953920

regulatory_elements/enhancer_dataset_subset.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7af8aef4a6c15a4be5d752f8a638f2ab88be42476f9da59732f892273d58078f
+size 11241394

regulatory_elements/promoter_dataset.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1ddaaa24bee906885033fb4f75097797c3aed3557b3fb09a202f50d4e99e3a2b
+size 33319573

regulatory_elements/promoter_dataset_subset.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4a5ccb2bf002dc504f6c40550acd5a99d2edfa587403d7cf23a1b965df74c51b
+size 7479637

variant_effect_causal_eqtl/All_Tissues.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d85df2b982e57866fb6efcd78dd0316c4005a0fea72322c3fe6c3188772558ae
+size 7081158

variant_effect_pathogenic/vep_pathogenic_coding.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e75b0b9c374e2cea972cb2c6246c7cc21c1c92e6b97cc597baf20b8d9462acb9
+size 6576847

variant_effect_pathogenic/vep_pathogenic_non_coding.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5ab457120792566dcfaeeaf06c9af32c19ddcff4d2f62fb034b95450622f9966
+size 369157498

variant_effect_pathogenic/vep_pathogenic_non_coding_subset.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:24127410937b99d41d0ad0b3b1e2e260b27ee20d00695caf6cbb48842a411a1f
+size 33142664