https://doi.org/10.5061/dryad.4j0zpc8mm

This dataset contains the gene count data used for finding biomarkers predicting corneal involvement in patients with acute infectious conjuntivitis, as described in the paper:

Seitzman GD, Prajna L, Prajna NV, et al. Biomarker Detection and Validation for Corneal Involvement in Patients With Acute Infectious Conjunctivitis. JAMA Ophthalmol. 2024;142(9):865–871. doi:10.1001/jamaophthalmol.2024.2891

List of files

biomarker_detection_count_data_md5-9b4b21f9.csv
biomarker_detection_sample_table_md5-7bd7fef2.csv
ENSG_ID2Name.txt.zip

Description of the data and file structure

Below is a brief description of each data file.

biomarker_detection_count_data_DESeq2norm_md5-9b4b21f9.csv

The CSV file biomarker_detection_count_data_DESeq2norm_md5-9b4b21f9.csv contains counts for human genes found in 58 conjunctival samples used in the study. RNA-Seq data was generated on an Illumina NovaSeq 6000 sequencing machine at the UCSF sequencing center. Sequencing reads were quality filtered using PriceSeqFilter, and aligned to the GRCh38 human genome assembly using HISAT2 (version 2.1.0). Abundance of genes was calculated using the default parameters in stringtie2 (version 1.3.4d). Annotation of transcripts was based on ENSEMBL GRCh38.87. The attached gene count matrix was then generated using the "prepDE.py" script according to the protocol found in the stringtie2 documentation.

• Rows: counts for 58,302 genes (identified by ENSG gene_id)
• Columns: one for each of 58 samples

biomarker_detection_sample_table_md5-7bd7fef2.csv

The CSV file biomarker_detection_sample_table_md5-7bd7fef2.csv contains metadata about the samples used, including the number of input reads for normalzing the counts to reads/million. It contains the following columns:

• Sample: unique, anonymized ID for each sample
• N_reads: number of sequencing read pairs
• Country: country of origin for each sample
• DESEq: part of training set used for DESeq2 dimensionality reduction
• Machine Learning: included in ML (all true, redundant column)
• RT-qPCR Validation: Real-time quantitative PCR was performed on this sample (Yes / No)
• Corneal Involvement: corneal involvement clinically detected (1 yes, 0 no)
• Sex: sex of patient
• Age: age of patient in years

ENSG_ID2Name.txt.zip

ENSG_ID2Name.txt.zip is a zip-compressed text file containing the mapping of ENSG IDs to gene names as they were used in the study.

Code/Software

Software used for determining gene counts:

• PRICE Sequence Filter (version 1.2)
• HISAT2 (version 2.1.0)
• stringtie2 (version 1.3.4d), and stringtie2's prepDE.py Python script