Longitudinal Olink Proteomics Data and Code

Repository for the paper: “Longitudinal proteomic profiling of dialysis patients with COVID-19 reveals markers of severity and predictors of death” by Gisby et al. doi: https://doi.org/10.7554/eLife.64827

Here, we have stored:

• A public version of the dataset used in the paper (data)
• An RMD script replicating our major analyses (scripts)
• A modified version of the OlinkAnalyze package containing functions used to fit models and generate plots (OlinkAnalyze_Modified, https://github.com/Olink-Proteomics/OlinkRPackage/tree/master/OlinkAnalyze)

Description of the Dataset

The dataset has been split into a series of four CSV files:

• plasma_npx_level.csv - Protein expression (NPX) data in long table format for the subcohort A (plasma)
• plasma_sample_level.csv - Sample level phenotypic data for the subcohort A
• serum_npx_level.csv - Protein expression (NPX) data in long table format for the subcohort B (serum)
• serum_sample_level.csv - Sample level phenotypic data for the subcohort B

Note that one individual ("C105") of the subcohort A was sampled both before and after their first positive swab, so the first sample is marked as "NEGATIVE" for Case_Control.

Protein Expression Data Features

For each of subcohorts A and B we provide the following columns:

Column Name	Data Type	Description
SampleID	Character	Unique identifier for samples
Individual_ID	Character	Unique identifier for individuals
UniProt	Character	UniProt ID for each protein
GeneID	Character	GeneID associated with each protein
Assay	Character	Name assigned to the protein by Olink
NPX	Numeric	Normalised protein expression (log2)
Panel	Character	The name of the panel each protein belongs to
Index	Integer	Unique sample identifier

Sample Level Data Features

The features of the sample level datasets include:

Column Name	Data Type	Description
SampleID	Character	Unique identifier for samples
Individual_ID	Character	Unique identifier for individuals
Plate_ID	Character	The assay plate to which the sample was assigned (1-5)
Case_Control	Character	Whether the individual was COVID POSITIVE or NEGATIVE at time of sampling. We sampled one individual ("C105") prior to their first positive swab, so they have both POSITIVE and NEGATIVE time points.
WHO_Severity_Peak	Character	The peak (WHO) severity for the patient over the disease course
WHO_Severity_Contemporaneous	Character	The (WHO) severity at time of sampling
Sex	Character	The individual's sex (M or F)
Ethnicity	Character	The individual's ethnicity (White, Black, South Asian, Asian (other) or Other)
Age	Character	Age in 20 year bins
Ever_Admitted	Logical	Whether the individual was admitted to hospital due to COVID at any point
Fatal_Disease	Logical	Whether the disease was fatal
Time_From_First_Symptoms	Integer	The number of days since the individual first experienced COVID symptoms at time of sampling
Time_Of_Death_From_First_Symptoms	Integer	The number of days between the individual's first experienced COVID symptoms and time of death
Time_From_First_Swab	Integer	The number of days since the individual's first positive swab was taken at time of sampling
Time_Of_Death_From_First_Swab	Integer	The number of days between the individual's first positive swab and time of death
WCC	Numeric	Individual's blood white cell count (x10^9/L) at time of sampling
Neutrophil	Numeric	Individual's blood neutrophil count (x10^9/L) at time of sampling
Monocyte	Numeric	Individual's blood monocyte count (x10^9/L) at time of sampling
Lymophocyte	Numeric	Individual's blood lymphocyte count (x10^9/L) at time of sampling
CRP	Numeric	Individual's blood C-Reactive Protein concentration (mg/L) at time of sampling
DDimer	Integer	Individual's blood D-Dimer concentration (ug/L) at time of sampling
Ferritin	Integer	Individual's blood ferritin concentration (ug/L) at time of sampling
Troponin	Integer	Individual's blood troponin concentration (ng/L) at time of sampling

Replicating the Analyses

We have provided data and R code for replicating the major analyses. The RMD script can be run (scripts/example_script.Rmd) to re-create the main results. Note that for the models we generated exact age was used, however we have only provided binned age in the public dataset. Therefore, exact results may differ slightly depending on the analysis.

The RMD script uses a modified OlinkAnalyze script to run the analyses and this must be installed prior to running. Code for this is also available in this repository if of interest.

License

This work is licensed under a Creative Commons Attribution 4.0 International License.