arXiv subjects

Sizhen Li

Publications and source records attributed to Sizhen Li.

4 recordsLinked to original sources

arXiv2024

Many-Shot In-Context Learning for Molecular Inverse Design

Large Language Models (LLMs) have demonstrated great performance in few-shot In-Context Learning (ICL) for a variety of generative and discriminative chemical design tasks. The newly expanded context windows of LLMs can further improve ICL capabilities for molecular inverse design and lead optimization. To take full advantage of these capabilities we developed a new semi-supervised learning method that overcomes the lack of experimental data available for many-shot ICL. Our approach involves iterative inclusion of LLM generated molecules with high predicted performance, along with experimental data. We further integrated our method in a multi-modal LLM which allows for the interactive modification of generated molecular structures using text instructions. As we show, the new method greatly improves upon existing ICL methods for molecular design while being accessible and easy to use for scientists.

cs.CL↗

arXiv2024

LinearAlifold: Linear-Time Consensus Structure Prediction for RNA Alignments

Predicting the consensus structure of a set of aligned RNA homologs is a convenient method to find conserved structures in an RNA genome, which has many applications including viral diagnostics and therapeutics. However, the most commonly used tool for this task, RNAalifold, is prohibitively slow for long sequences, due to a cubic scaling with the sequence length, taking over a day on 400 SARS-CoV-2 and SARS-related genomes (~30,000nt). We present LinearAlifold, a much faster alternative that scales linearly with both the sequence length and the number of sequences, based on our work LinearFold that folds a single RNA in linear time. Our work is orders of magnitude faster than RNAalifold (0.7 hours on the above 400 genomes, or ~36$\times$ speedup) and achieves higher accuracies when compared to a database of known structures. More interestingly, LinearAlifold's prediction on SARS-CoV-2 correlates well with experimentally determined structures, substantially outperforming RNAalifold. Finally, LinearAlifold supports two energy models (Vienna and BL*) and four modes: minimum free energy (MFE), maximum expected accuracy (MEA), ThreshKnot, and stochastic sampling, each of which takes under an hour for hundreds of SARS-CoV variants. Our resource is at: https://github.com/LinearFold/LinearAlifold (code) and http://linearfold.org/linear-alifold (server).

q-bio.BM↗

arXiv2023

LinearSankoff: Linear-time Simultaneous Folding and Alignment of RNA Homologs

The classical Sankoff algorithm for the simultaneous folding and alignment of homologous RNA sequences is highly influential, but it suffers from two major limitations in efficiency and modeling power. First, it takes $O(n^6)$ for two sequences where n is the average sequence length. Most implementations and variations reduce the runtime to $O(n^3)$ by restricting the alignment search space, but this is still too slow for long sequences such as full-length viral genomes. On the other hand, the Sankoff algorithm and all its existing implementations use a rather simplistic alignment model, which can result in poor alignment accuracy. To address these problems, we propose LinearSankoff, which seamlessly integrates the original Sankoff algorithm with a powerful Hidden Markov Model-based alignment module. This extension substantially improves alignment quality, which in turn benefits secondary structure prediction quality, confirmed over a diverse set of RNA families. LinearSankoff also applies beam search heuristics and the A$^\star$-like algorithm to achieve that runtime scales linearly with sequence length. LinearSankoff is the first linear-time algorithm for simultaneous folding and alignment, and the first such algorithm to scale to coronavirus genomes (n $\approx$ 30,000nt). It only takes 10 minutes for a pair of SARS-CoV-2 and SARS-related genomes, and outperforms previous work at identifying crucial conserved structures between the two genomes.

q-bio.BM↗

arXiv2022

LinearCoFold and LinearCoPartition: Linear-Time Algorithms for Secondary Structure Prediction of Interacting RNA molecules

Many ncRNAs function through RNA-RNA interactions. Fast and reliable RNA structure prediction with consideration of RNA-RNA interaction is useful. Some existing tools are less accurate due to omitting the competing of intermolecular and intramolecular base pairs, or focus more on predicting the binding region rather than predicting the complete secondary structure of two interacting strands. Vienna RNAcofold, which reduces the problem into the classical single sequence folding by concatenating two strands, scales in cubic time against the combined sequence length, and is slow for long sequences. To address these issues, we present LinearCoFold, which predicts the complete minimum free energy structure of two strands in linear runtime, and LinearCoPartition, which calculates the cofolding partition function and base pairing probabilities in linear runtime. LinearCoFold and LinearCoPartition follows the concatenation strategy of RNAcofold, but are orders of magnitude faster than RNAcofold. For example, on a sequence pair with combined length of 26,190 nt, LinearCoFold is 86.8x faster than RNAcofold MFE mode (0.6 minutes vs. 52.1 minutes), and LinearCoPartition is 642.3x faster than RNAcofold partition function mode (1.8 minutes vs. 1156.2 minutes). Different from the local algorithms, LinearCoFold and LinearCoPartition are global cofolding algorithms without restriction on base pair length. Surprisingly, LinearCoFold and LinearCoPartition's predictions have higher PPV and sensitivity of intermolecular base pairs. Furthermore, we apply LinearCoFold to predict the RNA-RNA interaction between SARS-CoV-2 gRNA and human U4 snRNA, which has been experimentally studied, and observe that LinearCoFold's prediction correlates better to the wet lab results.

q-bio.BM↗