Additional file 1 of Dissecting the small RNA code of inflammatory bowel disease

Additional file 1. Supplementary figures and tables Supplementary Figure S1. Length distribution of (A) tsRNAs, (B) ysRNAs, and (C) rsRNAs. The solid lines indicate the mean in RPM while the shaded areas indicate the 95% confidence interval. Supplementary Figure S2. Differentially expressed noncanonical sncRNA families in IBD in the Swedish cohort. Y-axis indicates the expression level in RPM. Using a linear model controlling for age and sex, 21 tsRNA/rsRNA/ysRNA families were found to be commonly upregulated (adjusted P < 0.05) in both the UC and CD samples compared to the HC group. Supplementary Figure S3. Five-fold cross-validation on the classification power of tsRNAs, rsRNAs, and ysRNAs in the Swedish cohort. (A) Histogram of mean AUC of 1,000 rounds of five-fold cross-validation. (B) Frequency of tsRNA/rsRNA/ysRNA families prioritized during the 1,000 rounds of cross-validation. In total, only 55 tsRNA/rsRNA/ysRNA families were prioritized at least once. The frequency of the sncRNA families within the 21-tsRNA/rsRNA/ysRNA signature ranges from 347 to 4,842 with a median of 2,515. Supplementary Figure S4. The top positively and negatively co-expressed ysRNA-gene pairs. The expression of ysRNA-RNY3 was positively correlated with that of gene SCX (left panel), while the expression of ysRNA-RNY3 was negatively correlated with that of geneSNX1 (right panel). The X-axes indicate the values of log2-transformed (RPM+1) of ysRNA-RNY3 while the Y-axes indicate the log2-transformed expression intensity of the individual genes. The correlation coefficients and P-values were calculated using Spearman’s rank correlation test. Supplementary Figure S5. ROC curves for the classification between the HC and UC samples and between the HC and CD samples in the Swedish cohort. Supplementary Figure S6. Correlation in t-statistic. For each cohort, differential expression analysis was performed between the HC and UC samples and between the HC and CD samples, respectively, using a linear model controlling for age and sex. Each dot represents one tsRNA/rsRNA/ysRNA family. Positive correlation in t-statistic computed by the linear model was observed between the comparison of HC vs. UC and the comparison of HC vs. CD. The red dots are the commonly dysregulated sncRNA families in both comparisons,i.e., 21 and 2 tsRNA/rsRNA/ysRNA families in the Swedish and German cohorts, respectively. The correlation coefficients andP-values were calculated using Spearman’s rank correlation test. Supplementary Figure S7. Differentially expressed miRNAs in IBD in the Swedish cohort. Y-axis indicates the miRNA expression in RPM. Using a linear model controlling for age and sex, 25 miRNAs were found to be commonly upregulated or downregulated (adjusted P < 0.05) in both the UC and CD samples compared to the HC group. Supplementary Table S1. Noncanonical sncRNA families Supplementary Table S2. Noncanonical sncRNA families commonly upregulated in the UC and CD samples in the Swedish cohort.