COVID-19 B-cell Epitope Prediction Dataset

Columns

• parent_protein_id: A unique identifier for the parent protein.
• protein_seq: The sequence of the parent protein.
• start_position: The starting position of the peptide sequence within the parent protein.
• end_position: The ending position of the peptide sequence within the parent protein.
• peptide_seq: The sequence of the peptide.
• chou_fasman: A peptide feature representing β-turn propensity.
• emini: A peptide feature indicating relative surface accessibility.
• kolaskar_tongaonkar: A peptide feature related to antigenicity.
• parker: A peptide feature indicating hydrophobicity.
• isoelectric_point: A protein feature describing its isoelectric point.
• aromacity: A protein feature indicating its aromaticity.
• hydrophobicity: A protein feature describing its hydrophobicity.
• stability: A protein feature indicating its stability.
• target: The antibody valence, which is the target value indicating whether a peptide exhibited antibody-inducing activity. This is a binary estimation (Positive or Negative).

Distribution

• input_bcell.csv is the primary training data, containing 14,387 rows, representing combinations of 14,362 unique peptides and 757 unique proteins. It includes all 14 columns.
  • parent_protein_id: 760 unique IDs.
  • protein_seq: 757 unique sequences.
  • start_position: Ranges from 1 to 3079, with a mean of 298.
  • end_position: Ranges from 6 to 3086, with a mean of 308.
  • peptide_seq: 14,362 unique sequences.
  • chou_fasman: Ranges from 0.53 to 1.55, with a mean of 0.99.
  • emini: Ranges from 0 to 27.2, with a mean of 1.06.
  • kolaskar_tongaonkar: Ranges from 0.84 to 1.25, with a mean of 1.02.
  • parker: Ranges from -9.03 to 9.12, with a mean of 1.77.
  • isoelectric_point: Ranges from 3.69 to 12.2, with a mean of 7.07.
  • aromaticity: Ranges from 0 to 0.18, with a mean of 0.08.
  • hydrophobicity (protein): Ranges from -1.97 to 1.27, with a mean of -0.41.
  • stability: Ranges from 5.45 to 137, with a mean of 43.7.
  • target: Binary values (0 for Negative, 1 for Positive), with 10,485 negative records and 3,902 positive records.
• input_sars.csv is also main training data, containing 520 rows.
• input_covid.csv is the target data and does not contain label information in its columns. All three datasets consist of protein and peptide information.

Usage

• Developing and testing algorithms for B-cell epitope prediction.
• Designing and developing vaccines that aim to induce antigen-specific antibody production.
• Conducting general medical data analysis, particularly in immunology and virology.
• Implementing and evaluating machine learning models for rapid vaccine development.

Coverage

License

Who Can Use It

• Data analysts: To explore and build predictive models for epitope identification.
• Vaccine researchers: For insights into antigen-antibody interactions and vaccine design.
• Bioinformaticians: To develop new computational methods for epitope prediction.
• Machine learning practitioners: To apply and advance machine learning techniques in biomedical research.

Dataset Name Suggestions

• COVID-19 B-cell Epitope Prediction Dataset
• SARS-CoV-2 Vaccine Epitope Data
• Immune Epitope Database for Vaccine Development
• Antigen Protein Epitope Prediction Data
• B-cell Antibody Induction Dataset

Attributes