Bias in RNA-seq Library Preparation: Current Challenges and Solutions

Table 4

The bias sources of major sequencing platforms.

Company	Platforms	Sequencing	Dominant bias type	Suggestion for improvement
Roche/454 Life Sciences	GS FLX Titanium XL+	Pyrosequencing	The bias of sequencing was introduced by PCR amplification prior to sequencing.	Reduction of the number of PCR cycles and use of DNA polymerases with even higher fidelity [84]
	GS FLX Titanium XLR70
	GS Junior
	HiSeq 2000
Illumina	Genome Analyzer IIx	Sequencing-by-synthesis with reversible terminator	Substitution type miscalls are the major source of bias.	Quality trimming (sickle) combined with error correction (BayesHammer) followed by read overlapping (PANDAseq) as the most suitable approach, reducing substitution biases [85]
	MiSeq
	SOLiD™ 4 system
	Ion PGM™ sequencer (318 chip)
Helicos BioSciences	HeliScope™ single molecule sequencer	Single-molecule sequencing	Biases were introduced by insertions and deletions.	If a low sequencing bias is needed, Illumina or SOLiD are often the best choices [86, 87]
Pacific Biosciences	PacBio RS	Single-molecule sequencing	High bias of raw single-pass data