TABLE 3.

Comparison of functionalities offered by UMIc, UMI-tools, and pRESTO.

Feature	UMIc	UMI-tools	pRESTO
Language	R	Python	Python
Input	fastq	bam	fastq
Align free	Yes	Not supported	Yes
Sequence length	Supports only same sequence length	Supports different sequence lengths	Supports different sequence lengths
Extract UMI	Based on the number of nucleotides on 5′	Based on the pattern of barcodes on 5′ and 3′	Based on the number of nucleotides on 5′
Paired data	Yes	Yes	Yes
UMI on R1 or R1 and R2	Yes	Yes	Yes
UMI correction	Through UMI and reads distance	Offers five methods (three of them network based, which use UMI distance and read counts) and one of them cutoff 1% of mean (number of reads/UMI)	In case of significant nucleotide diversity within UMI groups, divides the groups in subclusters
Data cleaning	Specification of min number of reads/UMI group	UMI quality filtering for a specified Phred score threshold	UMI quality filtering for a specified Phred score threshold and offers removal of highly variable UMI read groups
Deduplication	Creation of consensus sequence, using per base frequency and Phred scores	Selection of representative read, based on mapping coordinates and quality	Creation of consensus sequence, using per base Phred scores (optionally, a frequency and quality threshold that will assign an N to the position)
Output	fastq	bam	fastq