Table 2 Comparison of pipelines analysis steps
From: RATEmiRs: the rat atlas of tissue-specific and enriched miRNAs database
Analysis steps | Eli Lilly | NIEHS | Maastricht University |
|---|---|---|---|
Preprocessing | Trim adapters | Trim adapters | Trim adapters |
Quality filtering | Quality filtering | ||
Elimination of short reads | Elimination of short reads | Elimination of short reads | |
Alignment | miRDeep using miRBase v20 | BWA using miRBase v19 | Pre-alignment: miRDeep2 using Ensembl v5.0.73 |
Post-Alignment: PatMaN using predicted miRs and miRBase v20 | |||
Read count filtering | Yes (i.e. > 10 in at least one tissue/organ) | Yes (i.e. > 3 in at least one tissue/organ) | NA |
Scaling | NA | NA | According to # of loci |
Pre-Normalization | NA | TPMa | TMM |
Statistical method | Non-negative Matrix Factorization | Quasi-Poisson modeling | Percentage of total mapped read counts |
Comparison | One vs All | One vs Rest | One vs All |
Post-Normalization | Of basis (W) and coefficient (H) matrices | NA | NA |
Tissue-enriched thresholding | % of total expression (i.e. 60% in more than one tissue/organ) | p-value (i.e. < 0.05) | % of total expression (i.e. > 50%) |
Tissue/organ-specific thresholding | % of total expression (i.e. > 80% in one tissue/organ) | p-value and % of expression (i.e. <0.05 and > 90%) | % of total expression (i.e. > 90%) |
Data for display | TMM | TMM | TMM |