Sanejouand, Y.-H. | Search bioRxiv

bioRxiv2023

Are most human specific proteins encoded by long non-coding RNA?

By looking for a lack of homologues in a reference database of 27 well-annotated proteomes of primates and 52 well-annotated proteomes of other mammals, 170 putative human-specific proteins were identified. Among them, only 2 are known at the protein level and 23 at the transcript level, according to Uniprot. Though 21 of these 25 proteins are found encoded by an open reading frame of a long non-coding RNA, 60% of them are predicted to be at least 90% globular, with a single structural domain. However, there is a near complete lack of structural knowledge about these proteins, with no tridimensional structure presently available in the Protein Databank and a fair prediction for a single of them in the AlphaFold Protein Structure Database. Moreover, the knowledge about the function of these possibly key proteins remains scarce.

genomics↗

bioRxiv2017

PB-kPRED: Knowledge-Based Prediction Of Protein Backbone Conformation Using A Structural Alphabet

Libraries of structural prototypes that abstract protein local structures are known as structural alphabets and have proven to be very useful in various aspects of protein structure analyses and predictions. One such library, Protein Blocks (PBs), is composed of 16 standard 5-residues long structural prototypes. This form of analyzing proteins involves drafting its structure as a string of PBs. Thus, predicting the local structure of a protein in terms of protein blocks is a step towards the objective of predicting its 3-D structure. Here a new approach, kPred, is proposed towards this aim that is independent of the evolutionary information available. It involves (i) organizing the structural knowledge in the form of a database of pentapeptide fragments extracted from all protein structures in the PDB and (ii) apply a purely knowledge-based algorithm, not relying on secondary structure predictions or sequence alignment profiles, to scan this database and predict most probable backbone conformations for the protein local structures.\n\nBased on the strategy used for scanning the database, the method was able to achieve efficient mean Q16 accuracies between 40.8% and 66.3% for a non-redundant subset of the PDB filtered at 30% sequence identity cut-off. The impact of these scanning strategies on the prediction was evaluated and is discussed. A scoring function that gives a good estimate of the accuracy of prediction was further developed. This score estimates very well the accuracy of the algorithm (R2 of 0.82). An online version of the tool is provided freely for non-commercial usage at http://www.bo-protscience.fr/kpred/.

bioinformatics↗