Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].

Modelling sequential branching dynamics with a multivariate branching Gaussian process

Authors: Elvijs Sarkans, Sumon Ahmed, Magnus Rattray, Alexis Boukouvalas

TMLR 2023 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We examine the effectiveness of our approach on synthetic data and a single-cell RNA-Seq dataset from mouse haematopoietic stem cells (HSCs). Our approach ensures assignment consistency by design and achieves improved accuracy in branching time inference and assignment accuracy. 5 Experimental results We demonstrate our approach by fitting both MBGP and BGP models to noisy synthetic data. [...] We will also fit the MBGP model to real data, namely, single-cell RNA-seq data from mouse hematopoietic stem cells (Paul et al., 2015) and see how it resolves the cell label inconsistency problem whilst retaining biologically sensible fits.
Researcher Affiliation Collaboration Elvijs Sarkans EMAIL BIOS Health Sumon Ahmed EMAIL University of Dhaka Magnus Rattray EMAIL University of Manchester Alexis Boukouvalas EMAIL PROWLER.io
Pseudocode No The paper describes the model architecture and inference procedure using mathematical formulations and textual descriptions, but does not present any structured pseudocode or algorithm blocks.
Open Source Code No The paper does not contain any explicit statement about providing source code for the described methodology, nor does it provide a link to a code repository.
Open Datasets Yes We examine the effectiveness of our approach on synthetic data and a single-cell RNA-Seq dataset from mouse haematopoietic stem cells (HSCs). [...] We apply the MBGP model on the gene expression of mouse hematopoietic stem cells used in Paul et al. (2015). [...] Franziska Paul, Ya ara Arkin, Amir Giladi, Diego Adhemar Jaitin, Ephraim Kenigsberg, Hadas Keren-Shaul, Deborah Winter, David Lara-Astiaso, Meital Gury, Assaf Weiner, et al. Transcriptional heterogeneity and lineage commitment in myeloid progenitors. Cell, 163(7):1663 1677, 2015.
Dataset Splits No To speed up inference, we have randomly sub-sampled the data and have used 1000 cells for the inference and have used the sparse model with 60 inducing points. [...] We have used a sub-sample of 870 cells and 30 inducing points to speed up the model inference.
Hardware Specification Yes 1System configuration: Intel(R) Core(TM) i7-1065G7 CPU @ 1.30GHz 1.50 GHz, 16 GB RAM, 64-bit operating system, x64-based processor.
Software Dependencies No The paper mentions methods and algorithms like Monocle and Wishbone, but does not specify any software dependencies with version numbers for their own implementation of MBGP.
Experiment Setup Yes We initialise data point label probabilities for latent branches by sampling uniformly from [0.5, 1] for the true label. We then set the data point label prior probabilities to 0.5 for inputs in [0, 0.8] (that is, no prior knowledge) and then 0.8 for the true label for inputs in [0.8, 1], which we will refer to as the informative prior. [...] Unless explicitly stated otherwise, the fitting procedure is a multiple restart from 4 different branching points, namely, 0.2, 0.4, 0.6, 0.8 applied uniformly to all outputs. For each restart, we perform gradient descent along the evidence lower bound surface using L-BFGS-B.