Back to search

University of Crete (UOC)

Μελέτη της επιλεκτικής μακροαυτοφαγίας με μεθόδους βιοπληροφορικής και υπολογιστικής βιολογίας συστημάτων

Abstract

dc:description

In macroautophagy, several proteins interact with proteins of the ATG8 family through short linear motifs, mentioned in the literature as ATG8-family interacting motifs (AIMs) or LC3-interacting regions (LIRs). AIM/LIR motifs have been studied in a growing number of scientific papers in the last few years and these publications give significant information about macroautophagy in the eukaryotic cells. In ATG8, the LIR docking site (LDS) consists of two hydrophobic pockets (HP1 and HP2). A large number of potential LIR motifs are identified in proteins due to the small size of the motifs that follow the regular expression [WFY]xx[VLI]. The aromatic residue [WFY] binds to HP1 while the branched side chain aliphatic residue [VLI] binds to HP2. Using multiple experimental techniques, scientists have discovered several experimentally functional LIR motifs which interact with ATG8 homologs in multiple organisms. The available in silico LIR motif identification methodologies help to identify potential LIR motifs for experimental confirmation. Nevertheless, those methodologies give a significant number of false positive predictions. For the development of advanced computational tools which will contribute to the understanding and further experimental study of selective autophagy, the documentation and availability of data related with LIR motifs and the interaction with the ATG8 homologues are necessary.Using the Perl programming language the program “LIR-miner” was created. It uses text mining methods and it finds sentences and/or points referred to LIR motifs. Searching in PubMed Central using the keyword “autophagy”, 78798 open access publications were found. Those publications were analysed using the “LIR-miner” program to detect experimentally verified LIR-motifs. Using “LIR-miner”, direct search in PubMed, reverse search, i.e., papers citing key papers on LIR motifs or the software tools/databases for LIR motif prediction, 159 publications were found. In those publications 292 records with LIR-motifs were identified. For 9 of those motifs there is insufficient documentation for the experimental verification of the interaction of those motifs with the homologs of ATG8. The 82 motifs are experimentally verified non-functional motifs, 189 are experimentally verified functional motifs, 6 are Conditionally functional LIR-motifs and another 6 are Accessory LIR-motifs. A database was created and the data collected from the above process was added. At the same time, LIRcentral was created which is an online repository of experimentally verified LIR-motifs. LIRcentral is enriched with data from valid international databases. LIRcentral is freely available and open to the public. It is known that in addition to LIR-motif, the amino acids located in the region upstream or downstream of the LIR-motif also contribute to the interaction of the protein with ATG8. Some years ago, it was suggested that the PSSM score of the LIR-motif together with the two amino acids upstream of the motif can contribute to the identification of a functional LIR-motif. PSSM can describe structural/physicochemical characteristics upstream and downstream of a potential LIR-motif. Experiments were performed to identify the optimal upstream and downstream regions of the motif that contain useful information for distinguishing functional and non-functional motifs. The PSSMsearch software was used to calculate the PSSMs by choosing from several of the offered settings. The discrimination results for all possible combinations of upstream and downstream region lengths in the range [0,29] were then confirmed by cross-validation. It was found that the optimal choice for characterizing a functional LIR motif is using the PSSMs for 25 upstream amino acids, the LIR motif and 25 downstream amino acids. Two PSSM matrices were calculated. A PSSM matrix which uses all known functional LIR motifs and another PSSM matrix which uses all known non-functional motifs. For each known motif (functional or non-functional) two scores were calculated, a score based on the PSSM matrix of functional LIR-motifs and a score based on the PSSM matrix of the non-functional motifs. The scores were calculated summing the scores of the positions of 25 residues upstream the motif, four residues of the LIR-motif and 22 residues downstream the LIR-motif. The comparison of the two scores showed that for functional motifs, the score derived from the PSSM matrix of functional motifs tends to be greater than the score derived from the PSSM matrix of non-functional motifs. For non-functional motifs, the score derived from the PSSM matrix of non-functional motifs tends to be higher than the score derived from the PSSM matrix of functional motifs. It was also observed that the results improve by masking the aromatic and aliphatic residues of the LIR motif and recalculating the scores. The above result may help researchers to identify new potential functional LIR-motifs. A researcher wants to know the potential functional motif of a protein to validate it experimentally in the laboratory. To extract this information, a window of 51 amino acids was set which consist of the upstream+LIR+downstream (25+4+22=51) regions. Using this window, the amino acid sequence of a protein is scanned moving the window one amino acid each time. The windows of each protein are ranked in order starting from the higher score to the lower score. The 51-peptide containing the strong functional motif of the protein is expected to rank first with the highest PSSM functional motif score among the 51-peptides matching positions 26-29 with the motif [WFY]xx[VLI]. This procedure was encoded in the functional LIR motif prediction software LIRfinder, which is freely available as a web application: the user can either upload a file containing a protein in FASTA format or enter a protein in FASTA format in the relevant “text area” and submit the query.In our attempt to test the ability of PSSMs to successfully identify non-canonical motifs, we observed that this is not possible, since the aromatic or aliphatic residue may be absent. However, adding a “dummy” residue (Z) in place of the aromatic or aliphatic residue can give good results. This was helped by masking the aromatic and aliphatic residue when calculating the score of each 51-peptide since adding a Z at this position does not affect the calculated score of the 51-peptide. Extensive analysis was performed regarding (a) non-canonical motifs, (b) functional canonical motifs unidentified by the prediction model, (c) cases where the score of the functional motif is less than the score of a non-experimentally confirmed motif and (d) prediction results in the amino-terminal and carboxyl-terminal regions of proteins. Following the same procedure, proteins with experimentally verified LIR motifs, that were not contained in LIRcentral on the day of extraction of the data used in the experiments, were checked. This thesis led to (a) the development of computational tools for the semi-automated screening of scientific publications concerning the study of LIR motifs, (b) the methodical recording of experimentally verified motif data, (c) the creation of the LIRcentral service that provides easy access to the above data, and (d) the development and evaluation of the LIRfinder application that achieves improved prediction of functional LIR motifs. Given the central role of ATG8 family proteins in autophagy and other related cellular processes, there is great interest in the research community in studying their interactions with other protein molecules both in physiological conditions and in conditions related to forms of cellular stress or dysfunctions. For this reason, we believe that the above results will prove very useful in the effort to better understand the interactions that occur with the help of LIR motifs.

Degree

thesis:*
Grantor dc:publisher
University of Crete (UOC)
Year dc:date
2025

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Chatzichristofi, Agathangelos
  • Χατζηχριστοφή, Αγαθάγγελος

Subjects

dc:subject × 21

Rights

Language dc:language
gre

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/60111

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Chatzichristofi, Agathangelos; Χατζηχριστοφή, Αγαθάγγελος. Μελέτη της επιλεκτικής μακροαυτοφαγίας με μεθόδους βιοπληροφορικής και υπολογιστικής βιολογίας συστημάτων. University of Crete (UOC), 2025. http://hdl.handle.net/10442/hedi/60111