Hotspot Selective Preference of the Chimeric Sequences Formed in Multiple Displacement Amplification

Abstract

Multiple displacement amplification (MDA) is considered to be a conventional approach to comprehensive amplification from low input DNA. The chimeric reads generated in MDA lead to severe disruption in some studies, including those focusing on heterogeneity, structural variation, and genetic recombination. Meanwhile, the generation of by-products gives a new approach to gain insights into the reaction process of φ29 polymerase. Here, we analyzed 36.7 million chimeras and screened 196 billion chimeric hotspots in the human genome, as well as evaluating the hotspot selective preference of chimeras. No significant preference was captured in the distributions of chimeras and hotspots among chromosomes. Hotspots with overlaps for 12–13 nucleotides (nt) were most likely to be selected as templates in chimera generation. Meanwhile, a regularly selective preference was noticed in overlap GC content. The preferences in overlap length and GC content was shown to be pertinent to the sequence denaturation temperature, which pointed out the optimization direction for reducing chimeras. Distance preference between two segments of chimeras was 80–280 nt. The analysis is beneficial for reducing the chimeras in MDA, and the characterization of MDA chimeras is helpful in distinguishing MDA chimeras from chimeric sequences caused by disease.

ChimeraMiner: An Improved Chimeric Read Detection Pipeline and Its Application in Single Cell Sequencing

“…The results of the two pipelines demonstrated that the inverted chimeras are the dominant type of the single-end chimeras (average 91.36% by using ChimeraMiner, and average 88.69% by using the previous pipeline). The results were generally in concordance with the results of previous works [6,8,9], which endorsed that ChimeraMiner was accurate in chimera discovering.…”

“…In each single cell dataset, the inverted chimera was also the dominant type of the single-end chimera, which was coherent with the results of conventional MDA and previous works [6,8,9]. The consistent ratio between the inverted chimera and direct chimera, in all datasets, demonstrated that the application of ChimeraMiner in single cell MDA datasets was as effective as in conventional MDA datasets.…”

“…The limited amount of input DNA caused the increase of amplification duration, and the number of chimeras increased synchronously. Among single cell samples, inverted chimera was remaining the dominant type of the single-end chimera (Table 3), which is coherent with the results of conventional MDA and also the previous works [6,8,9]. The results demonstrated that applying ChimeraMiner in single cell MDA data was as affective as in conventional MDA datasets.…”

“…Chimeras resulted from alterative secondary structures, that can occur in the highly branched DNA, formed during the MDA processing [6], appearing as DNA rearrangements in amplified DNA (Figure 1a,b). The structure of chimeras is made up of two or more parts, which are located consecutively but adjacently on a chromosome (Figure 1c,d) [6,8,9]. In most bioinformatics analyses, chimeras can be filtered after mapping [5,10,11], but they still result in breaks in some researches, such as the ones aiming at structural variation (SV) detection [12,13,14,15,16,17], de novo genome assembly [18,19,20,21], and genetic recombination studies [22].…”

“…In previous works [8,9], authors divided chimeras into two categories: The insertion chimeras and the single-end chimeras. The insertion chimeras of the paired-end reads can be directly observed from the single-end mapped reads, based on the logical relation of mapping locations (+ or − strand).…”