Comparison of Sequence- Based Methods for sequence–structure alignment using the results of Picea abies L. genomics analysis

Mapping corresponding residues between the target sequence and template structure—a process referred to as ”sequence–structure alignment (SSA)”—is the first step in homology modeling. In a Picea abies L. genomics analysis, ۳۴% of the examined sequences had weak or no homology against PDB database . So we applied sequence-based methods for SSA on their results. Sensitivity and accuracy of alignments were compared to the result of DALI based on ۳D structure comparison method. Results revealed ۳ levels of sequence similarity based on decreasing di culty of identifying homology from sequence: ”midnight” zone (<۱۵% sequence identity), ”twilight” zone (۱۵–۲۵%), and ”daylight” zone (> ۲۵%). BLAST searches against PDB sequences led to an accurate alignment when sequence similarity was in daylight zone with no gaps. When there was a gap, alignment with gaps placement and boundary adjustments gave better results and MSA (multiple sequence alignment) ”MAFFT” helped refine target-template alignment. For ”twilight” zone and related to distant evolutionary relationships, pairwise sequence comparison was not su cient to reliably identify homology. So comparison of sequence profiles (PSI-BLAST) and HMMs ”hidden Markov model” (HMMER jackhmmer) with database sequences was done. Also, MSA method that uses additional information (predicted secondary structure, ۳D structural information) was used by ۳DCo↵ee. Evolutionary relationships that were too distant to be detected either by pairwise sequence or by profile or HMM–sequence comparisons were identified by methods that are based on profile–profile (COMA) or HMM–HMM (HHsearch) alignments. Results showed while BLAST did not detect some proteins, such as ARC۶ ”response to gibberellin,” COMA, HMMER, and PSI-BLAST had the best consistency with results derived from Dali. Depending on the type of protein family, alignment accuracy was di↵erent with each method, and HHsearch generated accurate and fast alignments. Protein families having more diverse and intermediate homologs resulted in more accurate alignments.

کلیدواژه ها:

Sequence-Based Methods ، Sequence–structure alignment ، Sequence similarity ، Picea abies L