A Comprehensive Approach to Sequence-oriented IsomiR annotation (CASMIR): demonstration with IsomiR profiling in colorectal neoplasia

Chung Wah Wu,John B Kisiel,Brian A Dukek,Douglas W Mahoney,David A Ahlquist,Jared M Evans,William R Taylor,Thomas C Smyrk,Jin Jen,Shengbing Huang,Tracy C Yab

doi:10.1186/s12864-018-4794-7

Chung Wah Wu, John B Kisiel + Show 9 more

Open Access

https://doi.org/10.1186/s12864-018-4794-7

Copy DOI

Abstract

BackgroundMicroRNA (miRNA) profiling is an important step in studying biological associations and identifying marker candidates. miRNA exists in isoforms, called isomiRs, which may exhibit distinct properties. With conventional profiling methods, limitations in assay and analysis platforms may compromise isomiR interrogation.ResultsWe introduce a comprehensive approach to sequence-oriented isomiR annotation (CASMIR) to allow unbiased identification of global isomiRs from small RNA sequencing data. In this approach, small RNA reads are maintained as independent sequences instead of being summarized under miRNA names. IsomiR features are identified through step-wise local alignment against canonical forms and precursor sequences. Through customizing the reference database, CASMIR is applicable to isomiR annotation across species. To demonstrate its application, we investigated isomiR profiles in normal and neoplastic human colorectal epithelia. We also ran miRDeep2, a popular miRNA analysis algorithm to validate isomiRs annotated by CASMIR. With CASMIR, specific and biologically relevant isomiR patterns could be identified. We note that specific isomiRs are often more abundant than their canonical forms. We identify isomiRs that are commonly up-regulated in both colorectal cancer and advanced adenoma, and illustrate advantages in targeting isomiRs as potential biomarkers over canonical forms.ConclusionsStudying miRNAs at the isomiR level could reveal new insight into miRNA biology and inform assay design for specific isomiRs. CASMIR facilitates comprehensive annotation of isomiR features in small RNA sequencing data for isomiR profiling and differential expression analysis.

Highlights

MicroRNA profiling is an important step in studying biological associations and identifying marker candidates. miRNA exists in isoforms, called isomiRs, which may exhibit distinct properties
IsomiR profiling of colorectal epithelia Small RNA sequencing on the 81 colorectal tissue samples generated 317 million total raw reads with an average read of 3.9 million per sample
Based on all 81 samples, canonical forms account for 31% of total miRNA reads, while non-canonical isomiRs account for 69%. 3′ isomiR is the most predominant form of isomiR, contributing 60% of all isomiRs (Fig. 3b), and 66% of all 3′ isomiRs fell into the addition form (Fig. 3c)

Summary

Introduction

MicroRNA (miRNA) profiling is an important step in studying biological associations and identifying marker candidates. miRNA exists in isoforms, called isomiRs, which may exhibit distinct properties. MicroRNA (miRNA) profiling is an important step in studying biological associations and identifying marker candidates. MiRNA exists in isoforms, called isomiRs, which may exhibit distinct properties. A list curated by the miRNA database (miRBase.org) defines the unique sequences referred to as the canonical forms. Accumulating evidence from deep sequencing suggests that miRNAs are heterogeneous in length and sequence, and the full array of isoforms is referred to as isomiRs [1]. The final products are 17–25 nucleotides in length, called mature miRNAs. Variants could originate from imprecise cleavage by Drosha or Dicer, generating sequences of different lengths [3]. IsomiRs could arise from post-transcriptional modifications initiated by nucleotidyl transferase, which predominantly adds specific nucleotides to pre-miRNA or mature miRNA ends [4,5,6].

Methods

Results

Discussion

Conclusion