In ssarda/genomeutils: Utilities for Genome Analysis

This package provides a set of helper tools that allow automating some common tasks one encounters during routine genome analysis.

Fast loading and writing of files and genome objects

Using the data.table package for performing these functions really fast.

Load from a file
Write to a file
Read .fasta/fastq to a Genome object
Write a Genome object to file
Read .fasta to a Proteome object
Write a Proteome object to file

Fetch gene attributes and functional analysis

Fetch the gene length (coding exons) for a list of genes
Fetch the GC content of the sequence for a list of genes
Perform Gene Ontology Enrichment analysis for a set of interesting genes

Normalization and Factorization methods for gene matrices

Min max normalization
Row median/deviation normalization
Sample specific normalization
Upper Quartile Normalization Gene counts to expression in 'Counts per million': (CPM)
Gene counts to expression in 'Transcripts per million': (TPM)
Gene counts to expression in 'Relative Log Expression': (RLE) as used in DESeq
Invertibility of matrix
Principal Component Analysis + 2D plots
Multi-Dimensional Scaling + 2D plots
Singular Value Decomposition + 2D plots

Pretty plotting functions

Modify heirarchical clustering to produce a plot colored by groups
Produce an MA plot (to identify differentially expressed genes, for instance)
Produce fancy heatmaps
Produce a smooth histogram by modifying base R plotting parameters

Hypothesis testing for matched groups of data

Includes ordering comparisons by significance

T-tests
F-tests
Significance of difference of means test
Significance of difference of variances test
Wilcoxon tests
Kolmogorov–Smirnov test

Maximum likelihood and Bayesian inference

Maximum likelihood estimation of Gaussian distribution parameters + AIC
Maximum likelihood estimation of Weibull distribution parameters + AIC
Implementation of binomial generalized linear model + AIC + BIC
Bayesian posterior estimation for a mixture of betas
Bayesian posterior estimation for binomial beta distribution

Machine learning

Support Vector Machine classifier
Naive Bayes Classifier
Random Forest Classifier
Linear Discriminant Analysis
Limma linear model for differential gene expression and computation of Residual sum of squares for downstream analysis

Variant Analysis (GWAS)

Carries out genome-wide association analysis using parallelized code to perform it really fast.
Input genotype and phenotype data and get significance measures by fitting a generalized linear model per SNP.

ssarda/genomeutils documentation built on May 30, 2019, 8:42 a.m.