Moreover, the soil temperature at the centre of the system for the horizontal alignment drops to lower values than for the other cases.
In contrast, the quality of RAF alignment drops dramatically when sequence similarity decreases; we conjecture that this is a consequence of the strong sequence-based heuristics in RAF.
The 70%_PERF_ALIGN threshold thus seems a good compromise between the number of conserved probesets and the fidelity of alignments, especially in cases of alignment drop due to intervening stop codons as found in the tomato GeneChip.
Requiring sequence alignment to both D. simulans and D. yakuba limits the data set to 11148 total gene alignments (2839 gene alignments are dropped) and it is likely that very rapidly evolving genes may not appear in the final data set.
For this dataset the alignment performance dropped considerably both for the pairwise and the multi-view reconstruction.
When performing assembly for a group of reads in a given cluster, reads that do not fit well in the overall alignment are dropped out and not included in the final formation of the "contigs", the consensus sequence representing a cluster.
The results are less satisfying with automatic alignments: concordance dropped to 60.5%, and there was a decrease in both DSC to 0.88 and 0.92 respectively.
Especially, when candidate and reference sequences share a lower fractional identity, alignment accuracy drops significantly.
For example, the twilight zone – the similarity range where alignment quality drops drastically – starts at 60%% for RNAs in comparison to 20%% for proteins.
