Hugging Face's logo

Datasets:
tatonettilab
/
onsides

Languages:
English
License:
Dataset card Data Studio Files
xet
 Community
Dataset Preview
Duplicate
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed because of a cast error
Error code:   DatasetGenerationCastError
Exception:    DatasetGenerationCastError
Message:      An error occurred while generating the dataset

All the data files must have the same columns, but at some point there are 6 new columns ({'effect_id', 'pred0', 'product_label_id', 'label_section', 'pred1', 'match_method'}) and 1 missing columns ({'ingredient_id'}).

This happened while the csv dataset builder was generating data using

hf://datasets/tatonettilab/onsides/product_adverse_effect.csv (at revision 05001835e4e53d3b4a7787f3c24c8ade77d85a80)

Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1831, in _prepare_split_single
                  writer.write_table(table)
                File "/usr/local/lib/python3.12/site-packages/datasets/arrow_writer.py", line 714, in write_table
                  pa_table = table_cast(pa_table, self._schema)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2272, in table_cast
                  return cast_table_to_schema(table, schema)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2218, in cast_table_to_schema
                  raise CastError(
              datasets.table.CastError: Couldn't cast
              product_label_id: int64
              effect_id: int64
              label_section: double
              effect_meddra_id: int64
              match_method: string
              pred0: double
              pred1: double
              -- schema metadata --
              pandas: '{"index_columns": [{"kind": "range", "name": null, "start": 0, "' + 1109
              to
              {'ingredient_id': Value('int64'), 'effect_meddra_id': Value('int64')}
              because column names don't match
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1339, in compute_config_parquet_and_info_response
                  parquet_operations = convert_to_parquet(builder)
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 972, in convert_to_parquet
                  builder.download_and_prepare(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 894, in download_and_prepare
                  self._download_and_prepare(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 970, in _download_and_prepare
                  self._prepare_split(split_generator, **prepare_split_kwargs)
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1702, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1833, in _prepare_split_single
                  raise DatasetGenerationCastError.from_cast_error(
              datasets.exceptions.DatasetGenerationCastError: An error occurred while generating the dataset
              
              All the data files must have the same columns, but at some point there are 6 new columns ({'effect_id', 'pred0', 'product_label_id', 'label_section', 'pred1', 'match_method'}) and 1 missing columns ({'ingredient_id'}).
              
              This happened while the csv dataset builder was generating data using
              
              hf://datasets/tatonettilab/onsides/product_adverse_effect.csv (at revision 05001835e4e53d3b4a7787f3c24c8ade77d85a80)
              
              Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

ingredient_id
int64
effect_meddra_id
int64
68,149
10,037,844
6,851
10,024,382
10,395
10,019,211
1,100,072
10,017,076
5,487
10,008,479
35,636
10,021,425
35,636
10,016,173
342,369
10,003,549
25,255
10,067,484
73,494
10,020,751
190,521
10,003,239
1,546,438
10,047,700
61,805
10,037,087
6,809
10,010,774
41,126
10,022,801
22,299
10,039,020
35,636
10,010,904
358,258
10,067,484
17,767
10,013,946
56,946
10,016,256
5,487
10,047,700
141,704
10,037,660
68,149
10,021,789
6,809
10,023,676
4,124
10,015,150
6,135
10,067,484
342,369
10,028,813
341,248
10,067,484
134,748
10,067,484
41,493
10,067,484
11,636
10,020,772
11,636
10,000,496
6,375
10,042,772
282,446
10,067,484
306,674
10,003,988
1,251
10,062,026
5,487
10,019,211
6,851
10,037,383
342,369
10,029,331
337,525
10,019,663
4,337
10,020,772
60,307
10,028,596
282,446
10,014,695
358,258
10,062,026
17,767
10,019,717
73,494
10,003,239
114,477
10,037,844
6,375
10,003,549
10,473
10,028,813
4,124
10,019,211
342,369
10,003,988
300,195
10,023,676
282,446
10,062,026
1,251
10,067,484
342,369
10,002,424
56,946
10,022,095
8,787
10,006,093
114,477
10,021,789
358,258
10,047,890
5,487
10,012,601
6,375
10,010,300
39,998
10,037,844
495,881
10,050,659
5,487
10,011,224
5,487
10,015,150
623,400
10,029,864
35,636
10,052,995
5,487
10,047,862
89,013
10,019,211
6,719
10,047,700
623,400
10,012,378
5,487
10,047,115
6,375
10,028,813
300,195
10,016,219
358,258
10,021,097
85,762
10,020,751
342,369
10,013,573
358,258
10,016,288
89,013
10,047,700
6,719
10,019,211
22,299
10,014,962
1,815
10,067,484
68,244
10,019,211
5,487
10,031,127
1,251
10,028,584
1,592,254
10,034,966
342,369
10,050,584
6,375
10,019,063
5,487
10,019,717
89,013
10,037,660
17,767
10,011,224
39,998
10,010,774
68,149
10,033,371
25,255
10,020,751
89,013
10,026,749
623,400
10,044,565
134,748
10,039,083
6,375
10,016,256
85,762
10,028,411
68,149
10,002,855
End of preview.

About

OnSIDES is a database of adverse drug events extracted from drug labels created by fine-tuning a PubMedBERT language model on 200 manually curated labels available from Denmer-Fushman et al.. This comprehensive database will be updated quarterly, and currently contains more than 3.6 million drug-ADE pairs for 2,793 drug ingredients extracted from 46,686 labels, processed from all of the labels available to download from DailyMed as of November 2023. Additionally, we now provide a number of complementary databases constructed using a similar method - OnSIDES-INTL, adverse drug events extracted from drug labels of other nations/regions (Japan, UK, EU), and OnSIDES-PED, adverse drug events specifically noted for pediatric patients in drug labels. We have recently released a preprint on medRxiv with a full description of the data, methods and analyses.

Model Accuracy

Our fine-tuned language model achieves an F1 score of 0.90, AUROC of 0.92, and AUPR of 0.95 at extracting effects from the ADVERSE REACTIONS section of the FDA drug label. For the BOXED WARNINGS section, the model achieves an F1 score of 0.71, AUROC of 0.85, and AUPR of 0.72. For the WARNINGS AND PRECUATIONS section, the model achieves an F1 score of 0.68, AUROC of 0.66, and AUPR of 0.68. Compared against the reference standard using the official evaluation script for TAC 2017, the model achieves a Micro-F1 score of 0.87 and a Macro-F1 of 0.85.

Citation

Tanaka Y, Chen HY, Belloni P, Gisladottir U, Kefeli J, Patterson J, Srinivasan A, Zietz M, Sirdeshmukh G, Berkowitz J, LaRow Brown K, Tatonetti NP. OnSIDES database: Extracting adverse drug events from drug labels using natural language processing models. Med. 2025 Mar 27:100642. doi: 10.1016/j.medj.2025.100642. PMID: 40179876.
Downloads last month
5
Total file size:
1.45 GB