Search

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ijms24054881/s1., Despite the diversity of liquid biopsy transcriptomic repertoire, numerous studies often exploit only a single RNA type signature for diagnostic biomarker potential. This frequently results in insufficient sensitivity and specificity necessary to reach diagnostic utility. Combinatorial biomarker approaches may offer a more reliable diagnosis. Here, we investigated the synergistic contributions of circRNA and mRNA signatures derived from blood platelets as biomarkers for lung cancer detection. We developed a comprehensive bioinformatics pipeline permitting an analysis of platelet- circRNA and mRNA derived from non-cancer individuals and lung cancer patients. An optimal selected signature is then used to generate the predictive classification model using machine learning algorithm. Using an individual signature of 21 circRNA and 28 mRNA, the predictive models reached an area under the curve (AUC) of 0.88 and 0.81, respectively. Importantly, combinatorial analysis including both types of RNAs resulted in an 8-target signature (6 mRNA and 2 circRNA), enhancing the differentiation of lung cancer from controls (AUC of 0.92). Additionally, we identified five biomarkers potentially specific for early-stage detection of lung cancer. Our proof-of-concept study presents the first multi-analyte-based approach for the analysis of platelets-derived biomarkers, providing a potential combinatorial diagnostic signature for lung cancer detection., European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie 765492.

The current versions of the microRNA databases MiRgeneDB, miRBase, and PmiREN contain annotations for a total of 358 different species. Public repositories, however, host small RNA sequencing data for over 800 species. This discrepancy implies that microRNA research is also very active in species that neither have an available high-quality genome assembly nor annotations for microRNAs or other types of noncoding genes. These cases are particularly challenging to analyze because reference sequences need to be collected from different sources and processed and formatted appropriately so that the dedicated small RNA analysis tools can make use of them. In this protocol we describe how small RNA sequencing data can be easily analyzed by means of a dockerized version of the well-established sRNAtoolbox/sRNAbench small RNA tools. We outline the analysis of two publicly available datasets to demonstrate basic aspects like the preparation of the local database, expression profiling, or differential expression analysis as well as more advanced features such as quantification of exogenous RNA content and data analysis in non-model species.

Additional file 2: Supplementary Figure 2. (A) Triple-coated or anti-PE beads were incubated for 10, 25 or 60 min in plasma (donor 6) spiked with HT29-CFSE EVs. Fluorescence signals were read on beads and specificity (S) determined. Highest average S/N ratios were obtained at 60min. Experiment performed in duplicate. (B) HT29-CFSE EVs were spiked in plasma (donor 6) diluted 1:2, 1:5, 1:10 or 1:20 with PBS, whilst maintaining a constant volume for IP with triple-coated beads. S/N ratios indicated that a similar recovery was achieved across dilutions. Experiment performed in duplicate.

Additional file 3: Supplementary Figure 3. mRNA profiles of CD61+ and CD9, CD63 or CD81+ EVs obtained from healthy donor plasma: exploratory data analysis and normalization. (A to D) Principal component analysis (PCA) was performed on unnormalized gene counts for exploratory data analysis. Principal components PC1 and PC2 were plotted on the X and Y axis, respectively. Different variables were evidenced: (A) Gene counts excluding internally-defined nCounter control genes, (B) Total read count per sample, (C) Sample groups, which in this case represented the antibodies immobilized on IP beads. Triple-coated (CD9, CD63 and CD81) or CD61-coated beads, (D) Due to a maximum number of 12 slots per nCounter experiment, samples were processed in different batches. Each batch represents one individual nCounter run. (E) Relative log expression (RLE) plots for visualization of the normalization performance with DESeq2 (left). Normalized count data was projected on PC1 and PC2 after PCA (right). Triple-coated and CD61 samples were represented in green and orange, respectively.

Additional file 6: Supplementary Table 1. Quantification of bead input and EV spikes used across IP experiments. We assured that beads always exceeded the number of EVs, but also avoided an overabundance of the former.

Additional file 1: Supplementary Figure 1. (A) Representative nFCM dot plot, showing a single-particle quantification of HEK293 EVs after CFSE staining. As determined after background correction (BC), nearly 88% of EVs incorporated the dye. (B) CD9 expression was determined by labelling HEK293 EVs with CD9-PE. Five independent experiments were plotted, resulting in a mean value of 42%. (C) HEK293-CFSE were spiked in plasma (donor 7) and IP conducted with triple-coated, anti-CD61 and anti-PE beads. Experiment performed in duplicate.

Additional file 4: Supplementary Figure 4. (A) Box plot evidencing raw counts obtained from healthy donor EV samples using both triple-coated or CD61 beads. A statistically significant difference between groups could be appreciated (p-value = 0.00053). (B) Box plot evidencing raw counts obtained from healthy donor and patient EV samples using CD61 beads. No statistically significant difference could be found between groups (p-value = 0.43).

Additional file 5: Supplementary Figure 5. EDA and normalization of gene expression data comparing two platelet-derived EV datasets obtained from healthy donors and early-stage NSCLC patients. (A to D) Principal component analysis (PCA) was performed on unnormalized gene counts for exploratory data analysis. Principal components PC1 and PC2 were plotted on the X and Y axis, respectively. Different variables are evidenced:(A) Gene counts excluding internally-defined nCounter control genes, (B) Total read count per sample, (C) Sample groups representing the two cohorts compared, healthy donors vs. early-stage NSCLC, (D) Due to a maximum number of 12 slots per nCounter experiment, samples were processed in different batches. Each batch represents one individual nCounter run. (E) Relative log expression (RLE) plots for visualization of the normalization performance with DESeq2 (left). Normalized count data was projected on PC1 and PC2 after PCA (right). Healthy donor and early-stage NSCLC patient samples were represented in green and orange, respectively.