Research progress of reduced amino acid alphabets in protein analysis and prediction

Liang, Yuchao; Yang, Siqi; Zheng, Lei; Wang, Hao; Zhou, Jian Zhong; Huang, Shenghui; Yang, Lei; Zuo, Yi

doi:10.1016/j.csbj.2022.07.001

Cited by 13 publications

(4 citation statements)

References 96 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This work demonstrated that the kmer AAR approach can be used as the basis for machine learning models capable of classifying sets of functionally related proteins with little sequence similarity. We concluded that the AAR allows greater flexibility in sequence representation and captures similarity between sequences that would not be detected via conventional methods, and these observations have been borne out in other studies as well ( Hauswedell et al , 2014 ; Liang et al , 2022 ). Thus, AAR creates simplified representations of proteins in the form of flexible feature vectors, which can then be used to construct machine learning models for functional prediction, explore sequence similarities in newly sequenced datasets (e.g.…”

Section: Introductionsupporting

confidence: 65%

Snekmer: a scalable pipeline for protein sequence fingerprinting based on amino acid recoding

Chang

Nelson

Jerger

et al. 2023

Bioinformatics Advances

View full text Add to dashboard Cite

Motivation The vast expansion of sequence data generated from single organisms and microbiomes has precipitated the need for faster and more sensitive methods to assess evolutionary and functional relationships between proteins. Representing proteins as sets of short peptide sequences (kmers) has been used for rapid, accurate classification of proteins into functional categories; however, this approach employs an exact-match methodology and thus may be limited in terms of sensitivity and coverage. We have previously used similarity groupings, based on the chemical properties of amino acids, to form reduced character sets and recode proteins. This amino acid recoding (AAR) approach simplifies the construction of protein representations in the form of kmer vectors, which can link sequences with distant sequence similarity and provide accurate classification of problematic protein families. Results Here we describe Snekmer, a software tool for recoding proteins into AAR kmer vectors and performing either (1) construction of supervised classification models trained on input protein families, or (2) clustering for de novo determination of protein families. We provide examples of the operation of the tool against a set of nitrogen cycling families originally collected using both standard hidden Markov models and a larger set of proteins from Uniprot and demonstrate that our method accurately differentiates these sequences in both operation modes. Availability and Implementation Snekmer is written in Python using Snakemake. Code and data used in this paper, along with tutorial notebooks, are available at http://github.com/PNNL-CompBio/Snekmer under an open source BSD-3 license. Supplementary information Supplementary data are available at Bioinformatics Advances online.

show abstract

Section: Introductionsupporting

confidence: 65%

Snekmer: a scalable pipeline for protein sequence fingerprinting based on amino acid recoding

Chang

Nelson

Jerger

et al. 2023

Bioinformatics Advances

View full text Add to dashboard Cite

show abstract

“…Additional to sequence- and structure-derived physicochemical properties, the presence/absence in protein sequences of certain k-mers was implemented as variable. To reduce sequence complexity, a reduced 7-letter aminoacid alphabet was used, as previously implemented in various machine learning methods applied to protein sequences 36,78 . A novel alphabet based on amino acid properties was designed, to define k-mers that may represent key motifs in protein physicochemistry (Supplementary Table 4b).…”

Section: Methodsmentioning

confidence: 99%

AMAPEC: accurate antimicrobial activity prediction for fungal effector proteins

Mesny,

Thomma

2024

Preprint

View full text Add to dashboard Cite

Fungi typically occur in environments where numerous and diverse other microbes occur as well, often resulting in fierce competition for nutrients and habitat. To support fungal fitness in these environments, they evolved various mechanisms that mediate direct antagonism towards niche competitors. Among these, the secretion of proteins with antimicrobial activities has been reported in fungi with diverse lifestyles. Recently, several plant-associated fungi were shown to rely on the secretion of antimicrobial effector proteins to antagonize certain members of plant hosts' microbiota and to successfully colonize plant tissues. Some of these effectors do not share homology with known antimicrobials and represent novel antibiotics. Accordingly, the occurrence and conservation of proteinaceous antimicrobials throughout the fungal tree of life remains enigmatic. Here we present a computational approach to annotate candidate antimicrobial effectors in fungal secretomes based on protein physicochemical properties. After curating a set of proteins that were experimentally verified to display antimicrobial activity and a set of proteins that lack such activity, we trained a machine learning classifier on properties of protein sequences and predicted structures. This predictor performs particularly well on fungal proteins (R2=0.89) according to our validations and is delivered as a software package named AMAPEC, dedicated to antimicrobial activity prediction for effector candidates. We subsequently used this novel software to predict antimicrobial effector catalogs in three phylogenetically distant fungi with distinct lifestyles, revealing relatively large catalogs of candidate antimicrobials for each of the three fungi, and suggesting a broad occurrence of such proteins throughout the fungal kingdom. Thus, AMAPEC is a unique method to uncover antimicrobials in fungal secretomes that are often sparsely functionally annotated, and may assist biological interpretations during omic analyses. It is freely available at https://github.com/fantin-mesny/amapec.

show abstract

“…Reduced amino acid alphabets are a common feature of protein aligners ( Liang et al 2022 ). A reduced amino acid alphabet is a small(er) alphabet, where groups of amino acids are each represented by a single letter.…”

Section: Methodsmentioning

confidence: 99%

Lambda3: homology search for protein, nucleotide, and bisulfite-converted sequences

Hauswedell,

Hetzel,

Gottlieb

et al. 2024

Bioinformatics

View full text Add to dashboard Cite

Motivation Local alignments of query sequences in large databases represent a core part of metagenomic studies and facilitate homology search. Following the development of NCBI Blast, many applications aimed to provide faster and equally sensitive local alignment frameworks. Most applications focus on protein alignments, while only few also facilitate DNA-based searches. None of the established programs allow searching DNA sequences from bisulfite sequencing experiments commonly used for DNA methylation profiling, for which specific alignment strategies need to be implemented. Results Here, we introduce Lambda3, a new version of the local alignment application Lambda. Lambda3 is the first solution that enables the search of protein, nucleotide as well as bisulfite-converted nucleotide query sequences. Its protein mode achieves comparable performance to that of the highly optimised protein alignment application Diamond, while the nucleotide mode consistently outperforms established local nucleotide aligners. Combined, Lambda3 presents a universal local alignment framework that enables fast and sensitive homology searches for a wide range of use cases. Availability and implementation Lambda3 is free and open source software publicly available at https://github.com/seqan/lambda/. Supplementary information Supplementary data are provided with this article.

show abstract

Research progress of reduced amino acid alphabets in protein analysis and prediction

Cited by 13 publications

References 96 publications

Snekmer: a scalable pipeline for protein sequence fingerprinting based on amino acid recoding

Snekmer: a scalable pipeline for protein sequence fingerprinting based on amino acid recoding

AMAPEC: accurate antimicrobial activity prediction for fungal effector proteins

Lambda3: homology search for protein, nucleotide, and bisulfite-converted sequences

Contact Info

Product

Resources

About