Multiple Sequence Alignment (MSA) plays a pivotal role in unveiling the evolutionary trajectories of protein families. The accuracy of protein structure predictions is often compromised for protein sequences that lack sufficient homologous information to construct high quality MSA. Although various methods have been proposed to generate virtual MSA under these conditions, they fall short in comprehensively capturing the intricate coevolutionary patterns within MSA or require guidance from external oracle models. Here we introduce MSAGPT, a novel approach to prompt protein structure predictions via MSA generative pretraining in the low MSA regime. MSAGPT employs a simple yet effective 2D evolutionary positional encoding scheme to model complex evolutionary patterns. Endowed by this, its flexible 1D MSA decoding framework facilitates zero or few shot learning. Moreover, we demonstrate that leveraging the feedback from AlphaFold2 can further enhance the model capacity via Rejective Fine tuning (RFT) and Reinforcement Learning from AF2 Feedback (RLAF). Extensive experiments confirm the efficacy of MSAGPT in generating faithful virtual MSA to enhance the structure prediction accuracy. The transfer learning capabilities also highlight its great potential for facilitating other protein tasks.
Protein-nucleic acid interactions play a very important role in a variety of biological activities. Accurate identification of nucleic acid-binding residues is a critical step in understanding the interaction mechanisms. Although many computationally based methods have been developed to predict nucleic acid-binding residues, challenges remain. In this study, a fast and accurate sequence-based method, called ESM-NBR, is proposed. In ESM-NBR, we first use the large protein language model ESM2 to extract discriminative biological properties feature representation from protein primary sequences; then, a multi-task deep learning model composed of stacked bidirectional long short-term memory (BiLSTM) and multi-layer perceptron (MLP) networks is employed to explore common and private information of DNA- and RNA-binding residues with ESM2 feature as input. Experimental results on benchmark data sets demonstrate that the prediction performance of ESM2 feature representation comprehensively outperforms evolutionary information-based hidden Markov model (HMM) features. Meanwhile, the ESM-NBR obtains the MCC values for DNA-binding residues prediction of 0.427 and 0.391 on two independent test sets, which are 18.61 and 10.45% higher than those of the second-best methods, respectively. Moreover, by completely discarding the time-cost multiple sequence alignment process, the prediction speed of ESM-NBR far exceeds that of existing methods (5.52s for a protein sequence of length 500, which is about 16 times faster than the second-fastest method). A user-friendly standalone package and the data of ESM-NBR are freely available for academic use at: https://github.com/wwzll123/ESM-NBR.
Generative protein language models are a natural way to design new proteins with desired functions. However, current models are either difficult to direct to produce a protein from a specific family of interest, or must be trained on a large multiple sequence alignment (MSA) from the specific family of interest, making them unable to benefit from transfer learning across families. To address this, we propose $\textbf{P}$r$\textbf{o}$tein $\textbf{E}$volutionary $\textbf{T}$ransformer (PoET), an autoregressive generative model of whole protein families that learns to generate sets of related proteins as sequences-of-sequences across tens of millions of natural protein sequence clusters. PoET can be used as a retrieval-augmented language model to generate and score arbitrary modifications conditioned on any protein family of interest, and can extrapolate from short context lengths to generalize well even for small families. This is enabled by a unique Transformer layer; we model tokens sequentially within sequences while attending between sequences order invariantly, allowing PoET to scale to context lengths beyond those used during training. In extensive experiments on deep mutational scanning datasets, we show that PoET outperforms existing protein language models and evolutionary sequence models for variant function prediction across proteins of all MSA depths. We also demonstrate PoET's ability to controllably generate new protein sequences.
With the exponential increase of the protein sequence databases over time, multiple-sequence alignment (MSA) methods, like PSI-BLAST, perform exhaustive and time-consuming database search to retrieve evolutionary information. The resulting position-specific scoring matrices (PSSMs) of such search engines represent a crucial input to many machine learning (ML) models in the field of bioinformatics and computational biology. A protein sequence is a collection of contiguous tokens or characters called amino acids (AAs). The analogy to natural language allowed us to exploit the recent advancements in the field of Natural Language Processing (NLP) and therefore transfer NLP state-of-the-art algorithms to bioinformatics. This research presents an Embedding Language Model (ELMo), converting a protein sequence to a numerical vector representation. While the original ELMo trained a 2-layer bidirectional Long Short-Term Memory (LSTMs) network following a two-path architecture, one for the forward and the second for the backward pass, by merging the idea of PSSMs with the concept of transfer-learning, this work introduces a novel bidirectional language model (bi-LM) with four times less free parameters and using rather a single path for both passes. The model was trained not only on predicting the next AA but also on the probability distribution of the next AA derived from similar, yet different sequences as summarized in a PSSM, simultaneously for multi-task learning, hence learning evolutionary information of protein sequences as well. The network architecture and the pre-trained model are made available as open source under the permissive MIT license on GitHub at https://github.com/issararab/PEvoLM.
The field of protein folding research has been greatly advanced by deep learning methods, with AlphaFold2 (AF2) demonstrating exceptional performance and atomic-level precision. As co-evolution is integral to protein structure prediction, AF2's accuracy is significantly influenced by the depth of multiple sequence alignment (MSA), which requires extensive exploration of a large protein database for similar sequences. However, not all protein sequences possess abundant homologous families, and consequently, AF2's performance can degrade on such queries, at times failing to produce meaningful results. To address this, we introduce a novel generative language model, MSA-Augmenter, which leverages protein-specific attention mechanisms and large-scale MSAs to generate useful, novel protein sequences not currently found in databases. These sequences supplement shallow MSAs, enhancing the accuracy of structural property predictions. Our experiments on CASP14 demonstrate that MSA-Augmenter can generate de novo sequences that retain co-evolutionary information from inferior MSAs, thereby improving protein structure prediction quality on top of strong AF2.
Profile hidden Markov models (pHMMs) are widely employed in various bioinformatics applications to identify similarities between biological sequences, such as DNA or protein sequences. In pHMMs, sequences are represented as graph structures. These probabilities are subsequently used to compute the similarity score between a sequence and a pHMM graph. The Baum-Welch algorithm, a prevalent and highly accurate method, utilizes these probabilities to optimize and compute similarity scores. However, the Baum-Welch algorithm is computationally intensive, and existing solutions offer either software-only or hardware-only approaches with fixed pHMM designs. We identify an urgent need for a flexible, high-performance, and energy-efficient HW/SW co-design to address the major inefficiencies in the Baum-Welch algorithm for pHMMs. We introduce ApHMM, the first flexible acceleration framework designed to significantly reduce both computational and energy overheads associated with the Baum-Welch algorithm for pHMMs. ApHMM tackles the major inefficiencies in the Baum-Welch algorithm by 1) designing flexible hardware to accommodate various pHMM designs, 2) exploiting predictable data dependency patterns through on-chip memory with memoization techniques, 3) rapidly filtering out negligible computations using a hardware-based filter, and 4) minimizing redundant computations. ApHMM achieves substantial speedups of 15.55x - 260.03x, 1.83x - 5.34x, and 27.97x when compared to CPU, GPU, and FPGA implementations of the Baum-Welch algorithm, respectively. ApHMM outperforms state-of-the-art CPU implementations in three key bioinformatics applications: 1) error correction, 2) protein family search, and 3) multiple sequence alignment, by 1.29x - 59.94x, 1.03x - 1.75x, and 1.03x - 1.95x, respectively, while improving their energy efficiency by 64.24x - 115.46x, 1.75x, 1.96x.
Large-scale Protein Language Models (PLMs) have improved performance in protein prediction tasks, ranging from 3D structure prediction to various function predictions. In particular, AlphaFold, a ground-breaking AI system, could potentially reshape structural biology. However, the utility of the PLM module in AlphaFold, Evoformer, has not been explored beyond structure prediction. In this paper, we investigate the representation ability of three popular PLMs: ESM-1b (single sequence), MSA-Transformer (multiple sequence alignment) and Evoformer (structural), with a special focus on Evoformer. Specifically, we aim to answer the following key questions: (i) Does the Evoformer trained as part of AlphaFold produce representations amenable to predicting protein function? (ii) If yes, can Evoformer replace ESM-1b and MSA-Transformer? (ii) How much do these PLMs rely on evolution-related protein data? In this regard, are they complementary to each other? We compare these models by empirical study along with new insights and conclusions. All code and datasets for reproducibility are available at https://github.com/elttaes/Revisiting-PLMs.
The development of data-dependent heuristics and representations for biological sequences that reflect their evolutionary distance is critical for large-scale biological research. However, popular machine learning approaches, based on continuous Euclidean spaces, have struggled with the discrete combinatorial formulation of the edit distance that models evolution and the hierarchical relationship that characterises real-world datasets. We present Neural Distance Embeddings (NeuroSEED), a general framework to embed sequences in geometric vector spaces, and illustrate the effectiveness of the hyperbolic space that captures the hierarchical structure and provides an average 22% reduction in embedding RMSE against the best competing geometry. The capacity of the framework and the significance of these improvements are then demonstrated devising supervised and unsupervised NeuroSEED approaches to multiple core tasks in bioinformatics. Benchmarked with common baselines, the proposed approaches display significant accuracy and/or runtime improvements on real-world datasets. As an example for hierarchical clustering, the proposed pretrained and from-scratch methods match the quality of competing baselines with 30x and 15x runtime reduction, respectively.
Increasing size of sequence databases caused by the development of high throughput sequencing, poses multiple alignment algorithms to face one of the greatest challenges yet. As we show, well-established techniques employed for increasing alignment quality, i.e., refinement and consistency, are ineffective when large protein families are of interest. We present QuickProbs 2, an algorithm for multiple sequence alignment. Based on probabilistic models, equipped with novel column-oriented refinement and selective consistency, it offers outstanding accuracy. When analysing hundreds of sequences, QuickProbs 2 is significantly better than Clustal Omega, the previous leader for processing numerous protein families. In the case of smaller sets, for which consistency-based methods are the best performing, QuickProbs 2 is also superior to the competitors. Due to computational scalability of selective consistency and utilisation of massively parallel architectures, presented algorithm is comparable to Clustal Omega in terms of execution time, and orders of magnitude faster than full consistency approaches, like MSAProbs or PicXAA. All these make QuickProbs 2 a useful tool for aligning families ranging from few, to hundreds of proteins. QuickProbs 2 is available at https://github.com/refresh-bio/QuickProbs.
Reinforcement learning algorithms require many samples when solving complex hierarchical tasks with sparse and delayed rewards. For such complex tasks, the recently proposed RUDDER uses reward redistribution to leverage steps in the Q-function that are associated with accomplishing sub-tasks. However, often only few episodes with high rewards are available as demonstrations since current exploration strategies cannot discover them in reasonable time. In this work, we introduce Align-RUDDER, which utilizes a profile model for reward redistribution that is obtained from multiple sequence alignment of demonstrations. Consequently, Align-RUDDER employs reward redistribution effectively and, thereby, drastically improves learning on few demonstrations. Align-RUDDER outperforms competitors on complex artificial tasks with delayed rewards and few demonstrations. On the Minecraft ObtainDiamond task, Align-RUDDER is able to mine a diamond, though not frequently. Code is available at https://github.com/ml-jku/align-rudder. YouTube: https://youtu.be/HO-_8ZUl-UY
Proteins are the major building blocks of life, and actuators of almost all chemical and biophysical events in living organisms. Their native structures in turn enable their biological functions which have a fundamental role in drug design. This motivates predicting the structure of a protein from its sequence of amino acids, a fundamental problem in computational biology. In this work, we demonstrate state-of-the-art protein structure prediction (PSP) results using embeddings and deep learning models for prediction of backbone atom distance matrices and torsion angles. We recover 3D coordinates of backbone atoms and reconstruct full atom protein by optimization. We create a new gold standard dataset of proteins which is comprehensive and easy to use. Our dataset consists of amino acid sequences, Q8 secondary structures, position specific scoring matrices, multiple sequence alignment co-evolutionary features, backbone atom distance matrices, torsion angles, and 3D coordinates. We evaluate the quality of our structure prediction by RMSD on the latest Critical Assessment of Techniques for Protein Structure Prediction (CASP) test data and demonstrate competitive results with the winning teams and AlphaFold in CASP13 and supersede the results of the winning teams in CASP12. We make our data, models, and code publicly available.
We present an extension of Felsenstein's algorithm to indel models defined on entire sequences, without the need to condition on one multiple alignment. The algorithm makes use of a generalization from probabilistic substitution matrices to weighted finite-state transducers. Our approach may equivalently be viewed as a probabilistic formulation of progressive multiple sequence alignment, using partial-order graphs to represent ensemble profiles of ancestral sequences. We present a hierarchical stochastic approximation technique which makes this algorithm tractable for alignment analyses of reasonable size.
Many biological questions, including the estimation of deep evolutionary histories and the detection of remote homology between protein sequences, rely upon multiple sequence alignments (MSAs) and phylogenetic trees of large datasets. However, accurate large-scale multiple sequence alignment is very difficult, especially when the dataset contains fragmentary sequences. We present UPP, an MSA method that uses a new machine learning technique - the Ensemble of Hidden Markov Models - that we propose here. UPP produces highly accurate alignments for both nucleotide and amino acid sequences, even on ultra-large datasets or datasets containing fragmentary sequences. UPP is available at https://github.com/smirarab/sepp.
A profile comparison method with position-specific scoring matrix (PSSM) is one of the most accurate alignment methods. Currently, cosine similarity and correlation coefficient are used as scoring functions of dynamic programming to calculate similarity between PSSMs. However, it is unclear that these functions are optimal for profile alignment methods. At least, by definition, these functions cannot capture non-linear relationships between profiles. Therefore, in this study, we attempted to discover a novel scoring function, which was more suitable for the profile comparison method than the existing ones. Firstly we implemented a new derivative free neural network by combining the conventional neural network with evolutionary strategy optimization method. Next, using the framework, the scoring function was optimized for aligning remote sequence pairs. Nepal, the pairwise profile aligner with the novel scoring function significantly improved both alignment sensitivity and precision, compared to aligners with the existing functions. Nepal improved alignment quality because of adaptation to remote sequence alignment and increasing the expressive power of similarity score. The novel scoring function can be realized using a simple matrix operation and easily incorporated into other aligners. With our scoring function, the performance of homology detection and/or multiple sequence alignment for remote homologous sequences would be further improved.