Software & Methods

In collaboration with the Pounds group, we developed the shinyCox R package to visualize survival functions predicted by Cox models. This software supports non-statisticians to interrogate Cox model results beyond traditional tabular summaries. We believe shinyCox will be impactful in maximizing our use of available resources and to drive advances in precision oncology and similar fields. 

Supporting paper: Selukar et al. Interactive Use of Cox Model–Predicted Survival Curves: An Application Using ACS10 Score and Age to Personalize Treatment of Pediatric AML. JCO Precis Oncol 9, e2500634(2025).

Our pipeline utilizes scRNA-seq reference and bulk transcriptomes to estimate cellular composition in the matched bulk proteomes. The expression of genes and proteins at either bulk level or cell type level can be integrated by the Angle-based Joint and Individual Variation Explained (AJIVE) framework. Meanwhile, MICSQTL can perform cell-type-specific quantitative trait loci (QTL) mapping to proteins or transcripts based on the input of bulk expression data and the estimated cellular composition per molecule type, without the need for single cell sequencing. We use matched transcriptome-proteome from human brain frontal cortex tissue samples to demonstrate the input and output of our tool.

Supporting paper: Pan Y, Wang X, Sun J, Liu C, Peng J, Li Q (2024). Multimodal joint deconvolution and integrative signature selection in proteomics. Communications Biology, 7(493). 

In many phase I trials, the design goal is to find the dose associated with a certain target toxicity rate. In some trials, the goal can be to find the dose with a certain weighted sum of rates of various toxicity grades. For others, the goal is to find the dose with a certain mean value of a continuous response. This package provides the setup and calculations needed to run a dose-finding trial with non-binary endpoints and performs simulations to assess design’s operating characteristics under various scenarios. Three dose finding designs are included in this package: unified phase I design (Ivanova et al. (2009) <doi:10.1111/j.1541-0420.2008.01045.x>), Quasi-CRM/Robust-Quasi-CRM (Yuan et al. (2007) <doi:10.1111/j.1541-0420.2006.00666.x>, Pan et al. (2014) <doi:10.1371/journal.pone.0098147>) and generalized BOIN design (Mu et al. (2018) <doi:10.1111/rssc.12263>). The toxicity endpoints can be handled with these functions including equivalent toxicity score (ETS), total toxicity burden (TTB), general continuous toxicity endpoints, with incorporating ordinal grade toxicity information into dose-finding procedure. These functions allow customization of design characteristics to vary sample size, cohort sizes, target dose-limiting toxicity (DLT) rates, discrete or continuous toxicity score, and incorporate safety and/or stopping rules.

We have proposed a versatile and efficient approach for Mendelian randomization analysis under different study designs. They can be random sampling design, extreme tails of the primary outcome of interest, or extreme tails of the risk factor that is the primary outcome of interest in the original study. Here, the risk factor is a continuous variable and can be gene expression or methylation data.

Supporting paper: Liyanage JSS*, Estepp J*, Srivastava K, Raskin S, Sheehan VA, Hankins J, Takemoto C, Li Y, Cui Y, Mori T, Burgess S, DeBaun M, Kang G#. A versatile and efficient novel approach for Mendelian randomization analysis with applications to assess the causal effect of fetal hemoglobin on anemia in sickle cell anemia. Mathematics 2022.

An R package Keyboard containing functions for the implementation and simulation of two phase I model-assisted maximum tolerated dose (MTD)-finding designs for single-agent and combination trials, and one biological dose (OBD)-finding phase I/II design.

Supporting paper: Chen Li, Hongying Sun, Cheng Cheng, Li Tang, Haitao Pan, A software tool for both the maximum tolerated dose and the optimal biological dose finding trials in early phase designs. Contemporary Clinical Trials Communications, 2022 

We propose a novel set-valued (SV) method to assess secondary trait genetic association studies and exposure-secondary outcome association studies using data collected under case-control study design (SV2bc) and extreme phenotype study design (STEPS). Here, secondary traits can be binary or continuous variables and can be associated with gene expression, gene methylation, etc.   

Supporting paper: Wenjian Bi, Yun Li, Matthew P Smeltzer, Guimin Gao, Shengli Zhao, Guolian Kang, STEPS: an efficient prospective likelihood approach to genetic association analyses of secondary traits in extreme phenotype sequencing, Biostatistics, Volume 21, Issue 1, January 2020, Pages 33–49,

Randomized clinical trials (RCTs) are considered the gold standard for clinical trials comparing treatment groups. However, historical control trials (HCTs) are an alternative to RCTs if randomization is not feasible because of ethical concerns, patient preference, limited patient populations, or regulatory acceptability. The primary benefit of HCTs is that all patients can receive the new treatment with historical data providing the information for the control arm. Therefore, HCTs are useful for studies with limited patient populations. Group sequential designs using Lan-DeMets error spending functions are proposed for historical control trials with time-to-event endpoints. Both O’Brien–Fleming and Gamma family types of sequential decision boundaries are derived based on sequential log-rank tests (Wu and Li, 2017).

Supporting paper: Wu J, Li Y. Group sequential design for historical control trials using error spending functions. J Biopharm Stat. 2020 Mar;30(2):351-363. doi: 10.1080/10543406.2019.1684305. Epub 2019 Nov 12. PMID: 31718458; PMCID: PMC7737423.

Department of Biostatistics
MS 768, Room R6030
St. Jude Children's Research Hospital