IQCELL: A platform for predicting the effect of gene perturbations on developmental trajectories using single-cell RNA-seq data

Abstract

The increasing availability of single-cell RNA-sequencing (scRNA-seq) data from various developmental systems provides the opportunity to infer gene regulatory networks (GRNs) directly from data. Herein we describe IQCELL , a platform to infer, simulate, and study executable logical GRNs directly from scRNA-seq data. Such executable GRNs provide an opportunity to inform fundamental hypotheses in developmental programs and help accelerate the design of stem cell-based technologies. We first describe the architecture of IQCELL. Next, we apply IQCELL to a scRNA-seq dataset of early mouse T-cell development and show that it can infer a priori over 75% of causal gene interactions previously reported via decades of research. We will also show that dynamic simulations of the derived GRN qualitatively recapitulate the effects of the known gene perturbations on the T-cell developmental trajectory. IQCELL is applicable to many developmental systems and offers a versatile tool to infer, simulate, and study GRNs in biological systems. (https://gitlab.com/stemcellbioengineering/iqcell)

OneSC: A computational platform for recapitulating cell state transitions

“…Then we computed precision and recall based on comparison of inferred GRNs to the ground truth GRNs ( Fig S4A-B ) against 13 other GRN inference methods (GENIE3 (Huynh-Thu et al ., 2010), GRNBoost2 (Moerman et al ., 2019), PPCOR (Kim, 2015), PyEpoch (E. Y. Su et al ., 2022), LEAP (Specht and Li, 2017), PIDC (Chan et al ., 2017), SCRIBE (Qiu et al ., 2020), SINCERITIES (Papili Gao et al ., 2018), SINGE (Deshpande et al ., 2022), SCODE (Matsumoto et al ., 2017), GRISLI (Aubin-Frankowski and Vert, 2020), GRNVBEM (Sanchez-Castillo et al ., 2018), IQCELL (Heydari et al ., 2021)). To be as fair as possible in this comparison, we selected edge weight thresholds that optimized F1 for those methods that produce edge weights (all methods except OneSC and IQCELL).The mean F1 score, which is the harmonic mean of precision and recall, of OneSC’s GRNs was 0.61 which was as good as or higher than other GRN inference methods applied to the same data ( Fig 2A ).…”

“…can the GRN model or simulate certain biological phenomenal such as cell differentiation?). Here, we define functional GRNs as those with the following two properties: 1) capable of generating rich dynamical behaviors that reflect biologically relevant steady states (Ye et al ., 2019; Guantes and Poyatos, 2008; Huang et al ., 2022; Heydari et al ., 2021; K. Su et al ., 2022) and 2) capable of generating perturbation predictions (Heydari et al ., 2021; K. Su et al ., 2022).…”

“…Boolean networks can satisfy these conditions, and have successfully been used to model a wide range of biological phenomenon such as embryonic stem cell self-renewal (Dunn et al ., 2019), transitions between pluripotent states (Yachie-Kinoshita et al ., 2018), T-cell development (Heydari et al ., 2021; Ildefonso and Finley, 2023), and Drosophila melanogaster segment polarity (Parmer et al ., 2022). Previously RE:IN was developed to synthesize Abstract Boolean Networks (a set of concrete Boolean networks) that are consistent with experimental constraints and was used to identify the transcription factors circuit for modelling naïve pluripotency (Dunn et al ., 2014).…”

“…Previously RE:IN was developed to synthesize Abstract Boolean Networks (a set of concrete Boolean networks) that are consistent with experimental constraints and was used to identify the transcription factors circuit for modelling naïve pluripotency (Dunn et al ., 2014). More recently, IQCELL used satisfiability modulo theory engine (Z3) (de Moura and Bjørner, 2008), a computational method that was also used in RE:IN (Dunn et al ., 2014) and Hamey et al (Hamey et al ., 2017), to identify Boolean logic functions that are constraint by the pseudo-time dynamics of binarized gene expressions (Heydari et al ., 2021). The authors have demonstrated IQCELL’s capabilities of constructing Boolean networks that model early mouse…”

scREMOTE: Using multimodal single cell data to predict regulatory gene relationships and to build a computational cell reprogramming model

“…In particular, this has facilitated the development of several computational methods for cell reprogramming, taking on a range of different approaches. These include differential expression [12,13,14], Boolean networks [15,16,17], dynamical systems [18,19], and regression [20]. However, these methods have significant limitations.…”