SNP Genomic Statistical Analysis Support
Budget: $250 – $750 USD
I am building predictive genomic models from a large-scale SNP array dataset and would like a skilled collaborator to take ownership of the statistical side of the pipeline. The raw data have already been collected; what I now need is a rigorous analysis workflow that turns those variants into clear, reproducible insights.
Before any modelling begins, the file will need a light round of quality assurance: filtering out low-quality calls, imputing missing genotypes, normalising across batches, and performing feature selection so that only informative loci move forward. Once this cleaned matrix is in place, the core assignment is to implement and interpret two complementary methods—regression analysis for association testing and principal component analysis for dimensionality reduction and structure correction.
Deliverables
. Most important point is to get binary classification of camels when predicted by the model with 90-95% accuracy.
. Mismatches are permitted only in the case of borderlines not the extreme misclassification.
• Reproducible scripts (R/Python or both) that perform the stated preprocessing steps
• Well-commented code for regression models and PCA, including visual summaries of key outputs
• A concise report or notebook explaining findings, parameter choices, and recommendations for the next modelling phase
Everything should run on standard open-source libraries (e.g., tidyverse, Bioconductor, scikit-learn) so the pipeline can be handed over or scaled easily. If you have alternative tool preferences I am flexible, as long as the results are transparent and reproducible.
I will provide the raw SNP files and existing metadata once we agree on the framework. Looking forward to collaborating with a bioinformatician who can move quickly from clean data to trustworthy statistics.
Before any modelling begins, the file will need a light round of quality assurance: filtering out low-quality calls, imputing missing genotypes, normalising across batches, and performing feature selection so that only informative loci move forward. Once this cleaned matrix is in place, the core assignment is to implement and interpret two complementary methods—regression analysis for association testing and principal component analysis for dimensionality reduction and structure correction.
Deliverables
. Most important point is to get binary classification of camels when predicted by the model with 90-95% accuracy.
. Mismatches are permitted only in the case of borderlines not the extreme misclassification.
• Reproducible scripts (R/Python or both) that perform the stated preprocessing steps
• Well-commented code for regression models and PCA, including visual summaries of key outputs
• A concise report or notebook explaining findings, parameter choices, and recommendations for the next modelling phase
Everything should run on standard open-source libraries (e.g., tidyverse, Bioconductor, scikit-learn) so the pipeline can be handed over or scaled easily. If you have alternative tool preferences I am flexible, as long as the results are transparent and reproducible.
I will provide the raw SNP files and existing metadata once we agree on the framework. Looking forward to collaborating with a bioinformatician who can move quickly from clean data to trustworthy statistics.