In the polymorphism detection step, each pool dataset was mapped against the reference sequence using SSAHASNP [http://www.sanger.ac.uk/Software/analysis/ssahaSNP/] with the -454 and the -NSQ parameters set to true.
Zebrafish datasets were mapped with Bowtie [ 52] against the zebrafish genome (danRer7) and a precompiled exon junction database.
All datasets were mapped using Bowtie (Langmead et al., 2009) using default parameters against, respectively, the hg19, mm9, rheMac2 and panTro2 genome assemblies obtained from the UCSC Genome Browser (Karolchik et al., 2008).
All datasets are mapped to hg18rpa.
The intronic miRNAs in Baskerville et al.'s dataset and Barad et al.'s dataset were mapped to the host genes in Su's dataset, respectively.
Quality trimmed reads from each of the three treatments were mapped against the "reference" dataset with CLC Genomics Workbench 5.1 as a short read aligner.
Cleaned reads for each individual sample for each species as described in Böhne et al. (2014) were mapped against our A. burtoni dataset of gene coding sequences.
To identify the genes that corresponded to the sequencing reads in each library of the present study, the purity-filtered high-quality reads were mapped against the reference transcriptome dataset, which was generated using Roche 454 pyrosequencing technology.
The sequenced and filtered putative small RNAs from our datasets, along with the miRBase entries for every mature, annotated miRNA, were mapped against the tammar genome using an ungapped short read aligner (see methods).
To study the effect of allelic differences between the sequence reads and the reference genome they were mapped against, we classified all read calls in the dataset as matching the reference allele or the non-reference allele.
All readseq reads from the publicly available datasets for S. cerevisiae (SRX000559-SRX000564) [ 31] and C. albicans (SRP002852) [ 32] were mapped against the respective custom reference sequences, allowing spliced and unspliced reads to be counted (as above).
