Noora68/PLiCat-dataset-0.1.0

• Paper: https://academic.oup.com/bib/article/26/6/bbaf665/8377155
• GitHub Repository: https://github.com/Noora68/PLiCat
• Online Demo: https://colab.research.google.com/drive/1wGSZsy7KyYoJf2PiXzP4SVLXonl-cWb9?usp=sharing

---

For the data processing methods for the training and test sets, please refer to github. The processing logic for the disease mutation dataset is as follows.

Data Acquisition and Annotation Pipeline for Disease Mutation Databases

Finally, we named the mutation dataset: disease_merged_35_500_filtered.csv

1. Data Acquisition from Multiple Disease Mutation Databases

1.1 ClinVar

• Download and Extract Data
  Download from:
  https://ftp.ncbi.nlm.nih.gov/pub/clinvar/tab_delimited/archive/
  Locate and download variant_summary.txt.gz.
  In a Jupyter Notebook, read, decompress, and save the file as ClinVar_data_summary.csv (7,131,811 records).

• Filter Lipid-Related Disease Data
  Filter data using a gene list related to lipid diseases and save as ClinVar_lipid.csv (1,045,883 records).

• Extract Key Fields
  Extract relevant columns and save as ClinVar_lipid_filtered.csv.

1.2 cBioPortal

• Study Identifiers:
  ucec_tcga_pan_can_atlas_2018, skcm_tcga_pan_can_atlas_2018, coadread_tcga_pan_can_atlas_2018, luad_tcga_pan_can_atlas_2018, stad_tcga_pan_can_atlas_2018, lusc_tcga_pan_can_atlas_2018, blca_tcga_pan_can_atlas_2018, brca_tcga_pan_can_atlas_2018, hnsc_tcga_pan_can_atlas_2018, cesc_tcga_pan_can_atlas_2018, gbm_tcga_pan_can_atlas_2018, lihc_tcga_pan_can_atlas_2018, ov_tcga_pan_can_atlas_2018, lgg_tcga_pan_can_atlas_2018, esca_tcga_pan_can_atlas_2018, prad_tcga_pan_can_atlas_2018, paad_tcga_pan_can_atlas_2018, kirp_tcga_pan_can_atlas_2018, kirc_tcga_pan_can_atlas_2018, sarc_tcga_pan_can_atlas_2018, thca_tcga_pan_can_atlas_2018, acc_tcga_pan_can_atlas_2018, ucs_tcga_pan_can_atlas_2018, laml_tcga_pan_can_atlas_2018, dlbc_tcga_pan_can_atlas_2018, thym_tcga_pan_can_atlas_2018, meso_tcga_pan_can_atlas_2018, kich_tcga_pan_can_atlas_2018, tgct_tcga_pan_can_atlas_2018, chol_tcga_pan_can_atlas_2018, pcpg_tcga_pan_can_atlas_2018, uvm_tcga_pan_can_atlas_2018, wt_target_2018_pub, all_phase2_target_2018_pub, aml_target_2018_pub, nbl_target_2018_pub, rt_target_2018_pub

• Download and Extract Data
  Download the tar.gz files for the above Study IDs from cBioPortal and save to ./pathogenic_dataset/cbioportal_downloads.
  Use a Jupyter Notebook to decompress and merge all files into a single CSV: cbioportal_data.csv (3,532,778 records).

• Filter Lipid-Related Disease Data
  Filter with gene list to get lipid-related diseases (374,223 records), save as cbioportal_lipid.csv.

• Extract Key Fields
  Save filtered important columns as cbioportal_lipid_filtered.csv.

---

2. Convert Genome Coordinates from GRCh37 to GRCh38 Using LiftOver

• Installation and Execution
  Install and run LiftOver on WSL command line (details in LiftOver.txt).

• Data Structure Inspection
  Review the output format after LiftOver conversion.

LiftOver is a genome coordinate conversion tool that translates genomic coordinates or annotation files (e.g., BED, VCF) between different reference genome versions due to coordinate system differences.
GRCh37 and GRCh38 are versions of the human genome reference sequences.

---

3. Annotate Data Using Ensembl VEP

• VEP Installation
  See installation instructions in Esembl VEP code.txt.

• Convert CSV to VCF for VEP Input
  Create input VCF files:

  • clinvar_vep_input.vcf
  • cbioportal_vep_input.vcf

• Run VEP Annotation
  Execute VEP annotation on WSL command line. Output files:

  • clinvar_vep_output.txt
  • cbioportal_vep_output.txt

• Post-Processing VEP Output
  Remove header comment lines and save as CSV:

  • clinvar_vep_head.csv
  • cbioportal_vep_head.csv

---

4. Obtain Protein Sequence Mutation Information

• Download Uniprot_sprot.fasta
  https://ftp.uniprot.org/pub/databases/uniprot/current_release/knowledgebase/complete/uniprot_sprot.fasta.gz

• Clean Synonymous and Non-Protein-Altering Mutations
  Remove synonymous mutations, duplicates, intronic/UTR/regulatory region mutations. Save cleaned files:

  • clinvar_vep_cleaned.csv
  • cbioportal_vep_cleaned.csv

• Remove Duplicates and Extract Sequences
  Using Uniprot_sprot.fasta, retrieve wild-type and mutant sequences, add columns "WT_seq" and "Mut_Seq". Save as:

  • clinvar_vep_with_sequences.csv
  • cbioportal_vep_with_sequences.csv

• De-duplication by “Amino_acids” and “SWISSPROT”
  Extract wild-type and mutant protein sequences, save in columns "WT_Seq" and "Mut_Seq".

---

5. Data Integration

• Merge Datasets
  Filter out rows with empty "WT_seq" and remove duplicate "WT_seq" rows.
  Vertically concatenate the two datasets, add a new column "Source".
  Save as:

  • clinvar_unique_WT_seq.csv
  • cbioportal_unique_WT_seq.csv
  • merged_unique_WT_seq_with_source.csv

• Filter Sequence Length
  Keep sequences with length between 35 and 500, save as merged_35_500.csv.

• Extract Useful Fields
  Save filtered key fields as merged_35_500_filtered.csv.

---

Citing this work

If you use the code or data in this package, please cite:

@Article{PLiCat2025,
  author  = {Feitong Dong,Jingrou Wu},
  journal = {Briefings in Bioinformatics},
  title   = {PLiCat:Decoding protein-lipid interactions by large language model},
  year    = {2025},
  volume  = {26},
  issue  = {6},
  doi     = {https://doi.org/10.1093/bib/bbaf665},
  URL     = {https://academic.oup.com/bib/article/26/6/bbaf665/8377155}
}