Recursion has published or presented multiple works of machine learning research using data generated from the Recursion OS, a platform built across diverse technologies that continuously expands one of the world’s largest proprietary biological and chemical datasets. Our goal is to help advance the field of machine learning research, methods development, and collaboration.


High-Resolution Genome-Wide Mapping of Chromosome-Arm-Scale Truncations Induced by CRISPR-Cas9 Editing

CRISPR-Cas9 editing is a scalable technology for mapping of biological pathways, but it has been reported to cause a variety of undesired large-scale structural changes to the genome. We performed an arrayed CRISPR-Cas9 scan of the genome in primary human cells, targeting 17,065 genes for knockout with 101,029 guides. High-dimensional phenomics reveals a “proximity bias” in which CRISPR knockouts bear unexpected phenotypic similarity to knockouts of biologically-unrelated genes on the same chromosome arm, recapitulating both canonical genome structure and structural variants. Transcriptomics connects proximity bias to chromosome-arm truncations. Analysis of published large-scale knockout and knockdown experiments confirms that this effect is general across cell types, labs, Cas9 delivery mechanisms, and assay modalities, and suggests proximity bias is caused by DNA double-strand-breaks with cell cycle control in a mediating role. Finally, we demonstrate a simple correction for large-scale CRISPR screens to mitigate this pervasive bias while preserving biological relationships.

Read the Paper

RxRx3: Phenomics Map of Biology

The combination of modern genetic perturbation techniques with high content screening has enabled genome-scale cell microscopy experiments that can be leveraged to construct maps of biology. These are built by processing microscopy images to produce readouts in unified and relatable representation space to capture known biological relationships and discover new ones. To further enable the scientific community to develop methods and insights from map-scale data, here we release RxRx3, the first ever public high-content screening dataset combining genome-scale CRISPR knockouts with multiple-concentration screening of small molecules (a set of FDA approved and commercially available bioactive compounds). The dataset contains 6-channel fluorescent microscopy images and associated deep learning embeddings from over 2.2 million wells that span 17,063 CRISPR knockouts and 1,674 compounds at 8 doses each. RxRx3 is one of the largest collections of cellular screening data, and as far as we know, the largest generated consistently via a common experimental protocol within a single laboratory. Our goal in releasing RxRx3 is to demonstrate the benefits of generating consistent data, enable the development of the machine learning methods on this scale of data and to foster research, methods development, and collaboration.

Read the PaperExplore the Dataset

Building and Benchmarking Representations of Life

Presented at the Learning Meaningful Representations of Life (LMRL) Workshop at NeurIPS 2022

The continued scaling of genetic perturbation technologies combined with high-dimensional assays (microscopy and RNA-sequencing) has enabled genome-scale reverse-genetics experiments that go beyond single-endpoint measurements of growth or lethality. Datasets emerging from these experiments can be combined to construct “maps of biology”, in which perturbation readouts are placed in unified, relatable embedding spaces to capture known biological relationships and discover new ones. Construction of maps involves many technical choices in both experimental and computational protocols, motivating the design of benchmark procedures by which to evaluate map quality in a systematic, unbiased manner. 

In this work, we propose a framework for the steps involved in map building and demonstrate key classes of benchmarks to assess the quality of a map. We describe univariate benchmarks assessing perturbation quality and multivariate benchmarks assessing recovery of known biological relationships from large-scale public data sources. We demonstrate the application and interpretation of these benchmarks through example maps of scRNA-seq and phenomic imaging data.

Read the PaperDownload the Poster

Watch the Presentation


Multi-Objective Generative Flow Networks

Presented at the AI for Accelerated Materials Design (AI4Mat) Workshop at NeurIPS 2022

In many applications of machine learning, like drug discovery and material design, the goal is to generate candidates that simultaneously maximize a set of objectives. As these objectives are often conflicting, there is no single candidate that simultaneously maximizes all objectives, but rather a set of Pareto-optimal candidates where one objective cannot be improved without worsening another. Moreover, in practice, these objectives are often under-specified, making the diversity of candidates a key consideration. The existing multi-objective optimization methods focus predominantly on covering the Pareto front, failing to capture diversity in the space of candidates. Motivated by the success of GFlowNets for generation of diverse candidates in a single objective setting, in this paper we consider Multi-Objective GFlowNets (MOGFNs). MOGFNs consist of a novel Conditional GFlowNet which models a family of single-objective sub-problems derived by decomposing the multi-objective optimization problem. Our work is the first to empirically demonstrate conditional GFlowNets. Through a series of experiments on synthetic and benchmark tasks, we empirically demonstrate that MOGFNs outperform existing methods in terms of Hypervolume, R2-distance and candidate diversity. We also demonstrate the effectiveness of MOGFNs over existing methods in active learning settings. Finally, we supplement our empirical results with a careful analysis of each component of MOGFNs.

Read the Paper

A Molecular Foundation Model for Drug Discovery

Presented at the Critical Assessment of Molecular Machine Learning (ML4Molecules) and the Learning Meaningful Representations of Life (LMRL) Workshops at NeurIPS 2022

Models that accurately predict properties based on chemical structure are valuable tools in drug discovery. However, for many properties, public and private training sets are typically small, and it is difficult for the models to generalize well outside of the training data. Recently, large language models have addressed this problem by using self-supervised pretraining on large unlabeled datasets, followed by fine-tuning on smaller, labeled datasets. In this paper, we report MolE, a molecular foundation model that adapts the DeBERTa architecture to be used on molecular graphs together with a two-step pretraining strategy. The first step of pretraining is a self-supervised approach focused on learning chemical structures, and the second step is a massive multi-task approach to learn biological information. We show that fine-tuning pretrained MolE achieves state-of-the-art results on 9 of the 22 ADMET tasks included in the Therapeutic Data Commons.

Read the PaperDownload the Poster

Watch the Presentation


Mapping Biology With a Unified Representation Space for Genomic and Chemical Perturbations to Enable Accelerated Drug Discovery

Presented at the Learning Meaningful Representations of Life (LMRL) Workshop at NeurIPS 2021

Biology is massively complex and highly networked, and as a result, modeling diseases and identifying safe and effective drugs to correct them is an enormous challenge. More than 90% of drugs that have been advanced to clinical trials fail before making it to the market. Over the last decade, we have witnessed exponential progress in diverse technical fields including genome editing, synthetic biology, robotics, automation, machine learning, artificial intelligence, and high-performance computation. Here we present Recursion’s application of these advancements to improve the scale and efficiency of drug discovery. We capture high-dimensional phenotypic readouts from human cells at a massive scale: 18,000 genes modeled by CRISPR knockout and 250,000 compound perturbations with a large number of replicates in multiple cell types. This enables us to unravel complex biological patterns and predict relationships between genes and compounds.Here, we demonstrate the feasibility of learning a representation of a broad range of biology, unified across genetic and chemical perturbations, that demonstrates power by recapitulating known pathways of genes and known drug mechanisms of action in diverse areas of biological function. Further, we show that the same learned representations can be used to direct search through chemical space to initiate discovery and structural optimization of new medicines.

Download the PosterWatch the Presentation
Stay informed about our COVID-19 study
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.