Twilight zone of protein sequence alignments.

By Burkhard Rost. 1999

Subfields: Computational Biology & Bioinformatics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Biology & Bioinformatics

Citations in the corpus

HH-suite3 for fast remote homology detection and deep protein annotation (2019, background): “HHblits can help to annotate proteins beyond the twilight zone [19].”
Evaluating Protein Transfer Learning with TAPE (2019, methodology): “We mainly rely on sequence identity, which measures the percentage of exact amino acid matches between aligned subsequences of proteins [19].”
Better prediction of functional effects for sequence variants (2015, neutral): “We retrieved all enzyme pairs with pairwise sequence identity >40% and HSSPvalues>0 [28-30].”
Computational Protein Design with Deep Learning Neural Networks (2018, background): “59 In a DNA repair enzyme 3-methyladenine DNA glycosylase, the probability that a random mutation can inactivate the protein was found to be 34%±6%, indicating a large proportion of mutations can be tolerated in this protein.”
ProteinNet: a standardized data set for machine learning of protein structure (2019, background): “y MSAs suitable for protein structure prediction. One MSA repository does exist (Joosten et al., 2011), but it appears out of date and is unsuitable for applications requiring deep homology searches (Rost, 1999). The substantial computational cost associated with generating MSAs may explain this sur”
Tighter Bounds for Structured Estimation (2008, background): “1 We note that the accuracy differences are most pronounced at the low percentage identity ranges, the ‘twilight zone’ regime where better alignment accuracy has far reaching consequences in many other computational biology applications [16].”
RSDB: representative protein sequence databases have high information content (2000, background): “It is well-known that above the 'twilight zone' of around 30–40% sequence identity in proteins it is often straightforward to search and align sequences (Chung and Subbiah, 1996; Doolittle, 1986; Rost, 1999).”
Alignment of protein sequences by their profiles (2004, unknown): “In contrast, despite the methodological advances listed above, alignments in the so-called “twilight zone” of less than 30% sequence identity still contain many errors (Rost 1999).”
Guidelines for the use and interpretation of assays for monitoring autophagy (3rd edition) (2012, unknown): “orthologs may share even less than 10% sequence identity at the amino acid sequence level.(1471) This is the case with autophagy subunits in protists(1472,1473) and with other universally conserved eukaryotic systems, as for example the nuclear pore complex.”
Analysing six types of protein-protein interfaces. (2003, methodology): “To reduce the bias, we compiled the largest possible non-redundant subset of PDB: no pair of proteins in that set had more than 25% identity over 100 aligned residues.(36) The non-redundant set included 1812 high-resolution structures.”
Secondary structure prediction with support vector machines (2003, methodology): “One source of this improvement is the continual expansion of the sequence databases (Rost, 2001b) and the other is the use of scoring matrices from iterated PSI-BLAST searches (Altschul et al., 1997) or hidden Markov model homology searches (Karplus et al., 1998), which recover information from…”
Iron homeostasis related genes in rice (2003, result): “All of the proteins with greater than 30% identity, with at least one of the reference proteins used in the searches, were regarded as functionally similar (homologous) to the reference proteins, receiving the same name (Orengo et al., 1994; Rost and Valencia, 1996; Rost, 1999; Todd et al., 2001).”
Recognizing protein substructure similarity using segmental threading. (2010, unknown): “, for the targets in ‘‘twilight-zone’’ (Rost, 1999).”
SNAP: predict effect of non-synonymous polymorphisms on function (2007, methodology): “Toward this end we aligned all experimentally annotated enzymes by pairwise BLAST (29), and selected only pairs that matched two criteria: (i) pairwise sequence identity 440%, (ii) HSSP-value 40 (30, 31 ).”