Biology subjects

Hassanin, A.

Publications and source records attributed to Hassanin, A..

7 recordsLinked to original sources

bioRxiv2026

Data-Centric Evaluation of Protein Function Prediction Pipelines

Performance estimates in protein function prediction depend not only on model choice but also on upstream decisions that define the learning problem. Using antioxidant protein classification as a controlled case study, we evaluated how dataset harmonisation, protein representation, redundancy control, and partitioning strategy affect protein machine learning pipelines. We integrated 18,804 records from 12 publicly available dataset entries into a curated consensus dataset of 4,193 protein sequences. One-hot encoding and six pretrained protein language model representations were evaluated as model inputs and as similarity spaces for redundancy reduction and distance-aware splitting. Representation choice substantially altered dataset geometry, retained dataset size, class balance, and downstream evaluation. At representation-specific p90 thresholds, one-hot encoding retained the complete dataset, whereas pretrained embeddings retained between 5% and 25% of sequences. Distance-aware partitioning reduced apparent performance relative to random splitting by up to 0.15 MCC before redundancy control, while this difference narrowed after similarity filtering. Selected configurations nevertheless maintained high performance under stricter evaluation, reaching an MCC of 0.84. These findings show that performance estimates should be interpreted as outcomes of complete data-centric workflows rather than isolated properties of predictive models.

bioinformatics↗

bioRxiv2026

Visualize, Explore, and Select: A protein Language Model-based Approach Enabling Navigation of Protein Sequence Space for Enzyme Discovery and Mining

The rapid expansion of protein sequence databases continues to outpace functional characterization, creating a persistent bottleneck in enzyme discovery and mining--particularly in large, heterogeneous, and sparsely annotated sequence spaces. This gap is amplified by visualization challenges and the lack of informed strategies for exploration, selection, and mining across sequence spaces. Here, we present an embedding-based workflow, implemented in a computational platform called SelectZyme, for alignment-free visualization and exploration of protein sequence space that combines protein language model (pLM) representations with dimensionality reduction and hierarchical density-based clustering. The approach links complementary visualizations of protein sequence space as low-dimensional landscapes, connectivity projections (minimum spanning trees), and dendrogram-based organization, enabling coherent interactive exploration and candidate selection across global context and local neighborhoods without relying on sequence-identity thresholds, EC numbers, conserved motifs, or predefined functional annotations. Across distinct case studies, we demonstrate that embedding-defined neighborhoods remain structurally conserved even when sequence identity falls within the twilight zone, and that coherent functional organization emerges also for modular protein segments in a fully unsupervised analysis. We also show how this workflow supports user friendly, interactive and scalable enzyme mining in a sparsely annotated, complex multi-family protein space surpassing > 100,000 sequences, enabling constrained candidate selection around experimentally validated anchors. By enabling interactive exploration across visualizations, supporting informed candidate selection, our workflow streamlines biocatalyst discovery and helps to bridge uncharacterized sequence-space to functional characterization campaigns - thus providing a broad starting point to downstream protein engineering, machine-learning-guided design cycles, and iterative experimental screening campaigns.

bioinformatics↗

bioRxiv2022

Identifying the most probable mammal reservoir hosts for Monkeypox virus based on ecological niche comparisons

BackgroundPrevious human cases or epidemics have suggested that Monkeypox virus (MPXV) can be transmitted through contacts with animals of African rainforests. Although MPXV has been identified in many mammal species, most are likely secondary hosts and the reservoir host has yet to be discovered. Methodology/Principal FindingsIn this study, we provide the full list of African mammal genera (and species) in which MPXV was previously detected and predict the geographic distributions of all species of these genera based on museum specimens and an ecological niche modelling (ENM) method. Then, we reconstruct the ecological niche of MPXV using georeferenced data on animal MPXV sequences and human index cases, and conduct overlap analyses with the ecological niches inferred for 99 mammal species in order to identify the most probable animal reservoir. Conclusions/SignificanceOur results show that the MPXV niche covers three African rainforests, the Congo Basin and Upper and Lower Guinean forests. The four mammal species showing the best niche overlap with MPXV are all arboreal rodents, including three squirrels, Funisciurus anerythrus, Funisciurus pyrropus, and Heliosciurus rufobrachium, and Graphiurus lorraineus. We conclude that the most probable MPXV reservoir is F. anerythrus based on two niche overlap metrics, the areas of higher probabilities of occurrence, and available data on MPXV detection. SummaryMonkeypox (MPX) is an emerging zoonotic disease, endemic to rainforests of West and Central Africa, which manifests as a fever, swollen lymph nodes, and fatigue, followed by a rash with macular lesions progressing from papules to vesicles, pustules and scabs, usually on the face, hands, and feet for two to four weeks. The case fatality rate ranges from 1-3% in West Africa to 5-10% in Central Africa. The disease has been reported in 10 African countries between 1970 and today, with an increasing number of cases over the last decades and several exportations outside the continent, the last one in 2022 resulting in an epidemic involving mostly men who have sex with men, with more than 80,000 cases detected worldwide. Although MPX virus has been identified in many mammal species, most are likely secondary hosts and the reservoir host has yet to be discovered. In this study, we compare the predicted geographic distribution (ecological niche) of the MPX virus with that of 99 mammal species, and conclude that the most probable MPX reservoir is Thomass rope squirrel (Funisciurus anerythrus), followed by three other arboreal rodents, including two squirrels (Funisciurus pyrropus and Heliosciurus rufobrachium) and the Lorrain dormouse (Graphiurus lorraineus).

ecology↗

bioRxiv2022

Draft genome of the lowland anoa (Bubalus depressicornis) and comparison with buffalo genome assemblies (Bovidae, Bubalina)

Genomic data for wild species of the genus Bubalus (Asian buffaloes) are still lacking while several whole genomes are currently available for domestic water buffaloes. To address this, we sequenced the genome of a wild endangered dwarf buffalo, the lowland anoa (Bubalus depressicornis), produced a draft genome assembly, and made comparison to published buffalo genomes. The lowland anoa genome assembly was 2.56 Gbp long and contained 103,135 contigs, the longest contig being 337.39 kbp long. N50 and L50 values were 38.73 kbp and 19.83 kbp, respectively, mean coverage was 44x and GC content was 41.74%. Two strategies were adopted to evaluate genome completeness: (i) determination of genomic features with de novo and homology-based predictions using annotations of chromosome-level genome assembly of the river buffalo, and (ii) employment of benchmarking against universal single-copy orthologs (BUSCO). Homology-based predictions identified 94.51% complete and 3.65% partial genomic features. De novo gene predictions identified 32,393 genes, representing 97.14% of the references annotated genes, whilst BUSCO search against the mammalian orthologues database identified 71.1% complete, 11.7% fragmented and 17.2% missing orthologues, indicating a good level of completeness for downstream analyses. Repeat analyses indicated that the lowland anoa genome contains 42.12% of repetitive regions. The genome assembly of the lowland anoa is expected to contribute to comparative genome analyses among bovid species.

genomics↗

bioRxiv2021

A novel SARS-CoV-2 related coronavirus in bats from Cambodia

Knowledge of the origin and reservoir of the coronavirus responsible for the ongoing COVID-19 pandemic is still fragmentary. To date, the closest relatives to SARS-CoV-2 have been detected in Rhinolophus bats sampled in the Yunnan province, China. Here we describe the identification of SARS-CoV-2 related coronaviruses in two Rhinolophus shameli bats sampled in Cambodia in 2010. Metagenomic sequencing identified nearly identical viruses sharing 92.6% nucleotide identity with SARS-CoV-2. Most genomic regions are closely related to SARS-CoV-2, with the exception of a small region corresponding to the spike N terminal domain. The discovery of these viruses in a bat species not found in China indicates that SARS-CoV-2 related viruses have a much wider geographic distribution than previously understood, and suggests that Southeast Asia represents a key area to consider in the ongoing search for the origins of SARS-CoV-2, and in future surveillance for coronaviruses.

microbiology↗

bioRxiv2020

Evolutionary history of Carnivora (Mammalia, Laurasiatheria) inferred from mitochondrial genomes

The order Carnivora, which currently includes 296 species classified into 16 families, is distributed across all continents. The phylogeny and the timing of diversifications are still a matter of debate. Here, complete mitochondrial genomes were analysed to reconstruct the phylogenetic relationships and to estimate divergence times among species of Carnivora. We assembled 51 new mitogenomes from 13 families, and aligned them with available mitogenomes by selecting only those showing more than 1% of nucleotide divergence and excluding those suspected to be of low-quality or from misidentified taxa. Our final alignment included 220 taxa representing 2,442 mitogenomes. Our analyses led to a robust resolution of suprafamilial and intrafamilial relationships. We identified 22 fossil calibration points to estimate a molecular timescale for carnivorans. According to our divergence time estimates, crown carnivorans appeared during or just after the Early Eocene Climatic Optimum; all major groups of Caniformia (Cynoidea/Arctoidea; Ursidae; Musteloidea/Pinnipedia) diverged from each other during the Eocene, while all major groups of Feliformia (Nandiniidae; Feloidea; Viverroidea) diversified more recently during the Oligocene, with a basal divergence of Nandinia at the Eocene/Oligocene transition; intrafamilial divergences occurred during the Miocene, except for the Procyonidae, as Potos separated from other genera during the Oligocene.

evolutionary biology↗

bioRxiv2020

The SARS-CoV-2-like virus found in captive pangolins from Guangdong should be better sequenced.

Viruses closely related to SARS-CoV-2, which is the virus responsible of the Covid-19 pandemic, were sequenced in several Sunda pangolins (Manis javanica) seized in the Guangdong and Guangxi provinces of China between 2017 and 20191-3. These viruses belong to two lineages: one from Guangdong (GD/P) and the other from Guangxi (GX/P). The GD/P viruses are particularly intriguing as the amino-acid sequence of the receptor binding domain of the spike protein is very similar to that of the human SARS-CoV-2 virus (97.4%)2. This characteristic suggests that GD/P viruses are capable of binding human ACE2 receptor and may therefore be able to mediate infection of human cells. Whereas all six GX/P genomes were deposited as annotated sequences in GenBank, none of the two GD/P genomes assembled in previous studies2,3 are currently available. To overcome this absence, I assembled these genomes from the Sequence Read Archive (SRA) data available for SARS-CoV-2-like viruses detected in five captive pangolins from Guangdong. I found the genome assemblies of GD/P virus of poor quality, having high levels of missing data. Additionally, unexpected reads in the Illumina sequencing data were identified. The GD/P2S dataset2 contains reads that are identical to SARS-CoV-2, suggesting either the coexistence of two SARS-CoV-2-like viruses in the same pangolin or contamination by the human virus. In the four other GD/P datasets1 many mitochondrial reads from pangolin were identified, as well as from three other species, namely, human, mouse and tiger. Importantly, I only identified three polymorphic nucleotide sites between the five GD/P sequences. Such low levels of polymorphism may reasonably be accounted for by sequencing errors alone, thus raising the possibility that the five pangolins seized in Guangdong in March 2019 were infected by the same virus strain, most probably during their captivity.

bioinformatics↗