Application suggestions of the FLt-Capture method
In our FLc-Capture experiment, we did not perform cDNA normalization to decrease the abundance of highly expressed transcripts, so the cDNA probe pools skew towards highly expressed genes. However, our FLc-Capture result showed that the capture depth of the obtained ORF and UTR sequences is not all related to the abundance of their cDNA probe (Fig. 5b); using unnormalized cDNA probes is still able to capture thousands of coding and noncoding loci. To obtain more uniform capture coverage across high and low-expressed transcripts, Puritz and Lotterhos (2017) used duplex-specific nuclease (DSN) treatment to prepare normalized cDNA probes for exome capture. They found that RNA sequencing coverage and exome sequencing coverage was still highly correlated (capture coverage will be higher for highly expressed genes) even using normalized cDNA probes. Therefore, it seems that the cDNA normalization is not an indispensable step. For projects focused on obtaining more phylogenetically informative loci for phylogenomic analysis, increasing the diversity of cDNA probes may be more effective than cDNA normalization. In such scenarios, we suggest to pool mRNAs extracted from multiple tissue types to create a high-diversity probe pool rather than using only liver mRNA as in our demonstration case.
Despite the demonstrated effectiveness of FLc-Capture from our snake case, one note should be considered before employing the method. Because the efficiency of sequence capture decreases with increased genetic distance between the probes and the targets, FLc-Capture might be less effective in large and highly divergent organism groups such as insects and other arthropods. In our demonstrating snake case, the maximal sequence difference between our probe species and outgroup species is about 15%, and we finally recovered 65% of the target ORF loci and 35% of the target UTR loci from these outgroup species, respectively. This threshold value (15% genetic difference) can be used as a starting point for other researchers to determine the phylogenetic depth of their FLc-Capture experiments. It has been shown that using DNA mixtures pooled from different representative species to prepare homemade probes is an effective strategy for sequence capture across large phylogenetic scales (Zhang et al., 2019). Therefore, if an investigator wants to apply the FLc-Capture method to a highly divergent organism group, it is possible to use several probe species that cover the entire phylogenetic span, mix their mRNAs to prepare a full-length cDNA probes, which can reduce the sequence divergence between probe and target. Of course, in such circumstances, the reference ORF and UTR sets should also be separated by different probe species, and the bioinformatics pipeline should be adjusted accordingly. This mixing strategy may allow for applying FLc-Capture across large phylogenetic scales but needs to be tested in the future.