In the study by Al-Khedher et al.6 the morphology of carbon nanotube structures, such as their curvature and alignment, was estimated using a neural network classifier.
The number of amino acid replacements per sequence position in the alignment was estimated using the JTT model [ 66].
For the paralinear distance analyses, the proportion of invariable sites in the considered alignment was estimated using likelihood and assuming a HKY85 + gamma + proportion of invariable sites model of DNA evolution.
A multiple sequence alignment was estimated using MUSCLE (Edgar, 2004) with MAXITERS set to 2, followed by the removal of identical sequences and the deletion of columns in which the seed had a gap.
For input, codon alignments were estimated using PAL2NAL [ 41] from a subset of sequences from the amino acid sequence alignment.
The best-fitting nucleotide substitution model for each codon-alignment was estimated using jModelTest 2.1.7 (Guindon and Gascuel 2003; Darriba et al. 2012) according to the corrected Akaike information criterion model.
Specifically, pairwise alignments of protein and mRNA sequences were estimated using a Smith-Waterman local alignment algorithm [ 79]; multiple sequence alignments were conducted using the ClustalW alignment algorithm [ 80].
Approximate alignment positions of mapped reads were estimated using the position of the spaced 20-mers, and all 60-mers existing in the read were stored in a hash table used by the de-novo assembler.
Bayesian phylogenetic trees were estimated using an alignment of the 11 full length DRA transcripts shown in Supplementary Figure S1 combined with the 33 DRA exon 2 sequences shown in Supplementary Figure S2.
