Biology subjects

Mofrad, M. R. K.

Publications and source records attributed to Mofrad, M. R. K..

10 recordsLinked to original sources

bioRxiv2025

Genome modeling and design across all domains of life with Evo 2

All of life encodes information with DNA. While tools for sequencing, synthesis, and editing of genomic code have transformed biological research, intelligently composing new biological systems would also require a deep understanding of the immense complexity encoded by genomes. We introduce Evo 2, a biological foundation model trained on 9.3 trillion DNA base pairs from a highly curated genomic atlas spanning all domains of life. We train Evo 2 with 7B and 40B parameters to have an unprecedented 1 million token context window with single-nucleotide resolution. Evo 2 learns from DNA sequence alone to accurately predict the functional impacts of genetic variation--from noncoding pathogenic mutations to clinically significant BRCA1 variants--without task-specific finetuning. Applying mechanistic interpretability analyses, we reveal that Evo 2 autonomously learns a breadth of biological features, including exon-intron boundaries, transcription factor binding sites, protein structural elements, and prophage genomic regions. Beyond its predictive capabilities, Evo 2 generates mitochondrial, prokaryotic, and eukaryotic sequences at genome scale with greater naturalness and coherence than previous methods. Guiding Evo 2 via inference-time search enables controllable generation of epigenomic structure, for which we demonstrate the first inference-time scaling results in biology. We make Evo 2 fully open, including model parameters, training code, inference code, and the OpenGenome2 dataset, to accelerate the exploration and design of biological complexity.

genomics↗

bioRxiv2023

Fine-tuning Protein Embeddings for GeneralizableAnnotation Propagation

A central goal of bioinformatics research is to understand proteins on a functional level, typically by extrapolating from experimental results with the protein sequence information. One strategy is to assume that proteins with similar sequences will also share function. This has the benefit of being interpretable; it gives a very clear idea of why a protein might have a particular function by comparing with the most similar reference example. However, direct machine learning classifiers now outperform pure sequence similarity methods in raw prediction ability. A hybrid method is to use pre-trained language models to create protein embeddings, and then indirectly predict protein function using their relative similarity. We find that fine-tuning an auxiliary objective on protein function indirectly improves these hy-brid methods, to the point that they are in some cases better than direct classifiers. Our empirical results demonstrate that interpretable protein comparison models can be developed using fine-tuning techniques, without cost, or even with some benefit, to overall performance. K-nearest neighbors (KNN) embedding-based models also offer free generalization to previously unknown classes, while continuing to outperform only pre-trained models, further demonstrating the potential of fine-tuned embeddings outside of direct classification. Significance StatementFor the purposes of biological understanding, or engineering, it is particularly important that protein annotation is done through interpretable means, such as comparison to existing proteins with known properties. Our work builds upon previous efforts to do so by embedding proteins into high-dimensional vectors with pre-trained protein language models. We demonstrate that further training of these models on protein function classification drastically improves their performance, while allowing for exactly the same similarity based classifications on both known and novel potential functions.

bioinformatics↗

bioRxiv2023

Force transmission and SUN-KASH higher-order assembly in the LINC complex models

The linker of the nucleoskeleton and cytoskeleton (LINC) complex comprises SUN (Sad-1 and UNC-84) and KASH (Klarsicht, ANC-1, SYNE homology) domain proteins, whose conserved interactions provide a physical coupling between the cytoskeleton and the nucleoskeleton, thereby mediating the transfer of physical forces across the nuclear envelope. The LINC complex can perform distinct cellular functions by pairing various KASH domain proteins with the same SUN domain protein. Recent studies have suggested a higher-order assembly of SUN and KASH instead of a more widely accepted linear trimer model for the LINC complex. In the present study, we use molecular dynamics simulations to investigate the mechanism of force transfer across the two proposed models of LINC complex assembly, namely the 3:3 linear trimer model and the 6:6 higher-order model. Employing steered molecular dynamics simulations with various structures using forces at different rates and directions, we examine the structural stability of the two models under various biologically relevant conditions. Our results suggest that both models can withstand and transfer significant levels of force while retaining their structural integrity. However, the force response of various SUN KASH assemblies depended on the force direction and pulling rates. Slower pulling rates resulted in higher mean square fluctuations of the 3:3 assembly compared to the fast pulling. Interestingly, the 6:6 assembly tends to provide an additional range of motion flexibility and might be more suitable for describing the interaction between SUN and KASH under compressive and shear forces. These findings offer insights into how the SUN and KASH proteins maintain the structural integrity of the nuclear membrane.

cell biology↗

bioRxiv2022

GO Bench: Shared-hub for Universal Benchmarking of Machine Learning-Based Protein Functional Annotations

MotivationGene annotation is the problem of mapping proteins to their functions represented as Gene Ontology terms, typically inferred based on the primary sequences. Gene annotation is a multi-label multi-class classification problem, which has generated growing interest for its uses in the characterization of millions of proteins with unknown functions. However, there is no standard GO dataset used for benchmarking the newly developed new machine learning models within the bioinformatics community. Thus, the significance of improvements for these models remains unclear. SummaryThe Gene Benchmarking database is the first effort to provide an easy-to-use and configurable hub for the learning and evaluation of gene annotation models. It provides easy access to pre-specified datasets and takes the non-trivial steps of preprocessing and filtering all data according to custom presets using a web interface. The GO bench web application can also be used to evaluate and display any trained model on leaderboards for annotation tasks. Availability and ImplementationThe GO Benchmarking dataset is freely available at llp.berkeley.edu/GO_bench/dataset_form, and code is available at http://github.com/amdson/GO_pipeline. Contactamdickson@berkeley.edu Supplementary informationSupplementary data are available at Bioinformatics online.

bioinformatics↗

bioRxiv2020

TripletProt: Deep Representation Learning of Proteins based on Siamese Networks

We introduce TripletProt, a new approach for protein representation learning based on the Siamese neural networks. We evaluate TripletProt comprehensively in protein functional annotation tasks including sub-cellular localization (14 categories) and gene ontology prediction (more than 2000 classes), which are both challenging multi-class multi-label classification machine learning problems. We compare the performance of TripletProt with the state-of-the-art approaches including recurrent language model-based approach (i.e., UniRep), as well as protein-protein interaction (PPI) network and sequence-based method (i.e., DeepGO). Our TripletProt showed an overall improvement of F1 score in the above mentioned comprehensive functional annotation tasks, solely relying on the PPI network. TripletProt and in general Siamese Network offer great potentials for the protein informatics tasks and can be widely applied to similar tasks.

bioinformatics↗

bioRxiv2020

The intriguing dynamics of chromatin folding and assembly

We investigate the dynamics of chromatin folding based on the "strings and binders" (SBS) model with molecular dynamics simulation. SBS model is a coarse-grained model considering a self-avoiding chain interacting with diffusive binders. By introducing transition among different categories of beads with specific transition cycles and transition probabilities, our model is capable of introducing different dynamics quantitatively during the folding process, thus capturing variety of phenomena related to chromatin dynamics. Firstly, roles of dynamics in the process of chromatin folding were examined. We discovered that there is a minimum gyration of chromatin under varying characteristic times of transition which indicates neither dramatically dynamic nor static folding process is optimal for chromatin to reach stable states with relatively low free energy. Secondly, it is noticeable that when beads transit from or into others in distinct dynamics, the equilibrium concentrations are distinct as well. As a consequence, the distribution of chromatin loop length is relevant to the dynamics of binders which can be modified by complex such as Wings apart-like protein homolog (Wapl) and SCC2/SCC4 cohesin loader complex (SCC2/SCC4). Finally, our model is able to reproduce contact matrices of both wild type HAP1 cell and {Delta}WAPL HAP1 cell obtained from Hi-C technology with a relatively high accuracy. Our model recapitulate the accumulating contacts at the corners of TADs and vanishing short-range contacts along the diagonal, manifesting the difference of chromatin structures before and after eliminating WAPL. STATEMENT OF SIGNIFICANCEOur model includes reciprocal transition among beads in SBS model to introduce different dynamics in chromatin folding process. Our model is able to examine the roles of dynamics in chromatin folding, reveal the loop length variation due to the concentration imbalance caused by distinct dynamics and reproduce contact matrices of both wild type and WAPL-deficient cells. Our research work provides a model to investigate the dynamics of chromatin folding quantitatively and displays its significance of revealing multiple experimental results using computational tools.

biophysics↗

bioRxiv2019

Kindlin Assists Talin to Promote Integrin Activation

Integrin IIb{beta}3 is a predominant type of integrin abundantly expressed on the surface of platelets and its activation regulates the process of thrombosis. Talin and kindlin are cytoplasmic proteins that bind to integrin and modulate its affinity for extracellular ligands. While the molecular details of talin-mediated integrin activation are known, the mechanism of kindlin involvement in this process remains elusive. Here, we demonstrate that the interplay between talin and kindlin promotes integrin activation. Our all-atomic molecular dynamics simulations on complete transmembrane and cytoplasmic domains of integrin IIb{beta}3, talin1 F2/F3 subdomains, and kindlin2 FERM domain in an explicit lipid-water environment over microsecond timescale, unraveled the role of kindlin as an enhancer of the talin interaction with the membrane proximal region of {beta}-integrin. The cooperation of kindlin with talin results in a complete disruption of salt bridges between R995 on IIb and D723/E726 on {beta}3. Furthermore, kindlin modifies the molecular mechanisms of inside-out activation by decreasing the crossing angle between transmembrane helices of integrin IIb-{beta}3, which eventually results in parallelization of integrin dimer. In addition, our control simulation featuring integrin in complex with kindlin reveals that kindlin binding is not sufficient for unclasping the inner membrane and outer membrane interactions of integrin dimer, thus ruling out the possibility of solitary action of kindlin in integrin activation.\n\nStatement of SignificanceUsing the newly solved crystal structure of kindlin, we investigated, for the first time, the molecular mechanism of kindlin-mediated integrin activation through simultaneous binding of talin and kindlin. We demonstrate in atomist details how kindlin cooperates with talin to promote the activation of integrin IIb-{beta}3.

biophysics↗

bioRxiv2018

Probabilistic variable-length segmentation of protein sequences for discriminative motif mining (DiMotif)and sequence embedding (ProtVecX)

In this paper, we present peptide-pair encoding (PPE), a general-purpose probabilistic segmentation of protein sequences into commonly occurring variable-length sub-sequences. The idea of PPE segmentation is inspired by the byte-pair encoding (BPE) text compression algorithm, which has recently gained popularity in subword neural machine translation. We modify this algorithm by adding a sampling framework allowing for multiple ways of segmenting a sequence. PPE segmentation steps can be learned over a large set of protein sequences (Swiss-Prot) or even a domain-specific dataset and then applied to a set of unseen sequences. This representation can be widely used as the input to any downstream machine learning tasks in protein bioinformatics. In particular, here, we introduce this representation through protein motif discovery and protein sequence embedding. (i) DiMotif: we present DiMotif as an alignment-free discriminative motif discovery method and evaluate the method for finding protein motifs in three different settings: (1) comparison of DiMotif with two existing approaches on 20 distinct motif discovery problems which are experimentally verified, (2) classification-based approach for the motifs extracted for integrins, integrin-binding proteins, and biofilm formation, and (3) in sequence pattern searching for nuclear localization signal. The DiMotif, in general, obtained high recall scores, while having a comparable F1 score with other methods in the discovery of experimentally verified motifs. Having high recall suggests that the DiMotif can be used for short-list creation for further experimental investigations on motifs. In the classification-based evaluation, the extracted motifs could reliably detect the integrins, integrin-binding, and biofilm formation-related proteins on a reserved set of sequences with high F1 scores. (ii) ProtVecX: we extend k-mer based protein vector (ProtVec) embedding to variable-length protein embedding using PPE sub-sequences. We show that the new method of embedding can marginally outperform ProtVec in enzyme prediction as well as toxin prediction tasks. In addition, we conclude that the embeddings are beneficial in protein classification tasks when they are combined with raw k-mer features.\n\nAvailabilityImplementations of our method will be available under the Apache 2 licence at http://llp.berkeley.edu/dimotif and http://llp.berkeley.edu/protvecx.

bioinformatics↗

bioRxiv2018

Nucleotide-pair encoding of 16S rRNA sequences for host phenotype and biomarker detection

Identifying combinations of taxa distinctive for microbiome-associated diseases is considered key to the establishment of diagnosis and therapy options in precision medicine and imposes high demands on accuracy of microbiome analysis techniques. We propose subsequence based 16S rRNA data analysis, as a new paradigm for microbiome phenotype classification and biomarker detection. This method and software called DiTaxa substitutes standard OTU-clustering or sequence-level analysis by segmenting 16S rRNA reads into the most frequent variable-length subsequences. These subsequences are then used as data representation for downstream phenotype prediction, biomarker detection and taxonomic analysis. Our proposed sequence segmentation called nucleotide-pair encoding (NPE) is an unsupervised data-driven segmentation inspired by Byte-pair encoding, a data compression algorithm. The identified subsequences represent commonly occurring sequence portions, which we found to be distinctive for taxa at varying evolutionary distances and highly informative for predicting host phenotypes. We compared the performance of DiTaxa to the state-of-the-art methods in disease phenotype prediction and biomarker detection, using human-associated 16S rRNA samples for periodontal disease, rheumatoid arthritis and inflammatory bowel diseases, as well as a synthetic benchmark dataset. DiTaxa identified 17 out of 29 taxa with confirmed links to periodontitis (recall= 0.59), relative to 3 out of 29 taxa (recall= 0.10) by the state-of-the-art method. On synthetic benchmark data, DiTaxa obtained full precision and recall in biomarker detection, compared to 0.91 and 0.90, respectively. In addition, machine-learning classifiers trained to predict host disease phenotypes based on the NPE representation performed competitively to the state-of-the art using OTUs or k-mers. For the rheumatoid arthritis dataset, DiTaxa substantially outperformed OTU features with a macro-F1 score of 0.76 compared to 0.65. Due to the alignment- and reference free nature, DiTaxa can efficiently run on large datasets. The full analysis of a large 16S rRNA dataset of 1359 samples required {approx}1.5 hours on 20 cores, while the standard pipeline needed {approx}6.5 hours in the same setting.\n\nAvailabilityAn implementation of our method called DiTaxa is available under the Apache 2 licence at http://llp.berkeley.edu/ditaxa.

bioinformatics↗

bioRxiv2018

MicroPheno: Predicting environments and host phenotypes from 16S rRNA gene sequencing using a k-mer based representation of shallow sub-samples

MotivationMicrobial communities play important roles in the function and maintenance of various biosystems, ranging from the human body to the environment. A major challenge in microbiome research is the classification of microbial communities of different environments or host phenotypes. The most common and cost-effective approach for such studies to date is 16S rRNA gene sequencing. Recent falls in sequencing costs have increased the demand for simple, efficient, and accurate methods for rapid detection or diagnosis with proved applications in medicine, agriculture, and forensic science. We describe a reference- and alignment-free approach for predicting environments and host phenotypes from 16S rRNA gene sequencing based on k-mer representations that benefits from a bootstrapping framework for investigating the sufficiency of shallow sub-samples. Deep learning methods as well as classical approaches were explored for predicting environments and host phenotypes.\n\nResultsk-mer distribution of shallow sub-samples outperformed the computationally costly Operational Taxonomic Unit (OTU) features in the tasks of body-site identification and Crohn's disease prediction. Aside from being more accurate, using k-mer features in shallow sub-samples allows (i) skipping computationally costly sequence alignments required in OTU-picking, and (ii) provided a proof of concept for the sufficiency of shallow and short-length 16S rRNA sequencing for phenotype prediction. In addition, k-mer features predicted representative 16S rRNA gene sequences of 18 ecological environments, and 5 organismal environments with high macro-F1 scores of 0.88 and 0.87. For large datasets, deep learning outperformed classical methods such as Random Forest and SVM.\n\nAvailabilityThe software and datasets are available at https://llp.berkeley.edu/micropheno.

bioinformatics↗