PLS regression - prediction of Intracellular PHA with Sudan-Black absorption

Published: 12 September 2024| Version 1 | DOI: 10.17632/frrxk4kzch.1
Contributors:
BISWANATH MAHANTY,

Description

This package contains codes for PLS modeling and analysis where in-vitro cellular absorption of SB is used to calculate intracellular protein content. The data set contains five measurements - free SB, cell-bound SB conc, intracellular protein analysis (Lowrey's assay at 650 nm), and cell OD. The package contains scripts to depict the SB partitioning, the PLS model development, the sensitivity of the PLS model, permitted predictor of importance, resilience to random error, method comparison using Bland-Altman Plot

Files

Steps to reproduce

Fig_1_Adsorption. mat, this file processes raw data on Sudan black free, Sudan black bound, protein, biomass OD, and PHA to generate two plots i.e. (a) Sudan black bound vs residual Sudan black, and (b) observed vs predicted Sudan black bound/free ratio. Fig_2_Calibration.mat, This file processes raw data on Sudan black free, Sudan black bound, protein, biomass OD, and PHA to generate three plots i.e. PHA vs (a) Sudan black bound, (b) OD at 650 nm, and (c) OD at 600 nm. Fig_3_Predimp_tolerence.mat, This file uses raw, calibration, validation, and model-predicted data to produce four outputs: regression between experimental and predicted PHA, factor loading PLS, predictor importance, and resilience to random measurement errors. Fig_S1_Sensitivity.mat, This file generates a sensitivity plot of three predictors—Sudan black bound, OD at 650 nm, and OD at 600 nm against intracellular PHA by taking the three predictors and one response value from mytable.mat as input. Fig_S3_4_Compare.mat, This file generates two plots validating the PLS model and determining MRE-based LOD for intracellular PHA by taking experimental response and PLS predicted response as input.

Institutions

Karunya University

Categories

Machine Learning, Error Resilience, Sensitivity Analysis, Mulitvariate Partial Least Squares Regression

Licence