Protoarray v5

ProtoArray^® v5.1 (PAH05251020, ThermoFisher, Waltham, MA, USA), containing replicates of >9000 full length human proteins for a total 23 232 spots per array (Supporting File 1). The proteins were expressed as N-terminal glutathione S-transferase (GST) fusions, expressed in using the Bac-to-Bac^® Baculovirus Expression System available from Invitrogen and affinity purified under native conditions to retain their proper conformation. Oligonucleotide detection probes labeled with the fluorophore FarRed with emission wavelength similar to Alexa Fluor^® 647 recommended for the reader, or antibodies conjugated with FITC, were used to detect RCA products and GST tag-specific anti-GST antibodies (DyLight^® 550). The arrays were scanned using the CapitalBio LuxScan HT24 at two different wavelengths: F635 (red) for the FarRed detection probe and F635 (532) for the FITC-labeled detection probe or for anti-GST-conjugated DayLight^® 550, serving to detect total protein amounts per spot. Data acquisition, alignment and image processing were performed using the GenePix^® Pro microarray (v6.1) software. Statistical analysis of protein array data was based on log-transformed intensities and statistical software ‘R’. The histogram was created using ‘ggplot2 package R’.

Probing and scanning of the protein arrays (ProtoArray® v5.0 PAH0525020, Life Technology) was conducted according to Invitrogen’s protocol for Immune Response BioMarker Profiling, using the recommended detection reagent (Alexa Fluor® 647 Goat Anti-Human IgG A21445, Invitrogen) and blocking buffer (Blocking Buffer Kit PA055, Invitrogen). The arrays were probed with sera at a dilution of 1:2000. Arrays were scanned using a GenePix 4000B microarray scanner, and the GenePix® Pro microarray (v6.1) software was used for alignment and data acquisition.

Human protein microarray (ProtoArray^© V5.0), were purchased from Invitrogen, Carlsbad, CA. Each ProtoArray protein array contains around 9,000 GST-tagged full-length human proteins in duplicates. These full-length N-terminal GST fusion proteins are expressed using Baculovirus expression system and were purified under non-denaturing conditions to maintain the protein integrity and function. The purified proteins are then printed on an ultrathin layer of nitrocellulose coated glass slides under temperature and humidity-controlled environment. Each block on a ProtoArray slide contains positive (Alexa Flour Ab, Human IgG, Anti-human IgG, and V5 control protein) and negative (Buffer, BSA, and GST) control spotted in duplicates (Figure 1B).

A commercially available human protein microarray (Protoarray v5.0; Invitrogen) spotted with >9,000 human full-length proteins purified from a baculovirus-based expression system previously successfully employed to identify ARHGAP26 as the target antigen of anti-Ca was probed with the patient serum according to the manufacturer’s instructions as described [30 (link)].

We downloaded multiple protein expression profiles from the GEO database (https://www.ncbi.nlm.nih.gov/geo/), including GSE29654, GSE62283, GSE74763. These datasets were generated by the Invitrogen ProtoArray V5.0 platform. Since there were some duplicate samples in these three datasets, we kept the duplicate samples with the earlier upload time. We also removed samples from the same institution but with a sample size of less than 3. Finally, the number of samples in each category, as shown in Table 1.
The format of raw data is GPR, we use the R package PPA to load it, then normalized the data with the robust linear model (RLM) method, which is the standard intra-slice method capable of ignoring the effect of isolated points, allowing a good fit of the regression line. Finally, common probes were extracted and expression profiles were merged. We standardized the data by the following method, for each gene in each sample, calculated the ratio of that gene to the total gene expression in the sample and multiplied the ratio by 1 million as the final expression value. The formula is as follows, where the data is a two-dimensional table with i rows and j columns, the row stores the protein, j represents the sample, x^ij represents the expression value of the ith protein of the jth sample, and x^ij’ represents the standardized data. $x_{ij}^{\text{'}}=\left(\frac{x_{ij}}{{{\displaystyle \sum }}_{i=1}^n{x}_{ij}}\right)\ast 1000000$