DiffRAC: A flexible R framework for comparing response proportions in high-throughput sequencing count data
Differential Ratio Analysis in Count data (DiffRAC) is a R framework to compare sequencing count ratios using a customized model matrix and DESeq2, from an experimental design, a formula and a read count tables.
An example analysis can be found at ./examples.
The main function. Infers the change in the response ratios using the customized model matrix and DESeq2, from the experimental design, the read count matrix, and the formula provided by the user
DiffRAC_res <- DiffRAC(formula, design, counts_num, counts_denom, mode="condition", bias=1, optimizeBias=F)The design data frame. Each row is one sample, and each column is an experimental variable. Sample names should be indicated as row.names, and the experimental variables as column names. Sample names have to match the column names in the count matrices (but not necessarily in the same order). The variables may be factors or numerical, and the user needs to make sure that the data has the intended class. There must be no NA values in the design. A minimum of two replicates per condition (or per combination of conditions) for the variable of interest should be used. For example:
| samples | V1 | V2 |
|---|---|---|
| cond1_rep1 | 1 | 0 |
| cond1_rep2 | 1 | 0 |
| cond2_rep1 | 1 | 1 |
| cond2_rep2 | 1 | 1 |
| cond3_rep1 | 0 | 0 |
| cond3_rep2 | 0 | 0 |
| cond4_rep1 | 0 | 1 |
| cond4_rep1 | 0 | 1 |
Another valid example:
| samples | cell_type |
|---|---|
| s1 | cellLineA |
| s2 | cellLineA |
| s3 | cellLineB |
| s4 | cellLineB |
Which is equivalent to:
| samples | cellLineA |
|---|---|
| s1 | 0 |
| s2 | 0 |
| s3 | 1 |
| s4 | 1 |
Please avoid the use of "-" and "." in the sample names. Names should also not start with a number.
A formula indicating the relationship between the predictor and outcome variables. The variable names must be the same as in the design matrix. For example:
~ V1 + V2 + V1:V2
The count matrix or count data frame for the numerator type. The rows and columns must match and be in the same order as the count_denom table. Column names should match the sample names in the design matrix (but not necessarily with the same order). Names should not start with a number. The row names must contain gene IDs.
| Gene_ID | cond1_rep1 | cond1_rep2 | cond2_rep1 | cond2_rep2 | cond3_rep1 | cond3_rep2 | cond4_rep1 | cond4_rep2 |
|---|---|---|---|---|---|---|---|---|
| 22746 | 29 | 3 | 27 | 2 | 47 | 3 | 37 | 5 |
| 235623 | 122 | 92 | 90 | 18 | 299 | 45 | 454 | 6 |
| 238690 | 18 | 14 | 6 | 8 | 71 | 22 | 60 | 34 |
| 330369 | 5 | 35 | 4 | 17 | 149 | 55 | 276 | 23 |
| ... | ... | ... | ... | ... | ... | ... | ... | ... |
The count matrix or count data frame for the denominator type, similar to count_num.
Either "condition" or "sample". Optionally, for small sample sizes, a sample-specific analysis can be performed, using the sample option, instead of a condition-specific investigation. This will significantly increase the run time. The default is mode="condition"
The "bias" constant. Optionally, the bias term can be supplied by the user. The default is bias=1
Optionally, an optimization can be performed to obtain the bias term, using optimizeBias=T. The default is optimizeBias=F.
DiffRAC returns a list with three elements:
model_mat: The customized model matrix created by DiffRACcounts: The counts used by DiffRACdds: The DESeq dds object containing the differential estimates.
The user can then use their own contrasts and follow the DESeq2 documentation to get differential estimates (as a log2 fold-change) and identify differential events. Please note that in case of factor variables, the variable names will be different from the inputed design and formula to the resulting dds object. For example, for a variable V1 with levels 0 and 1, a column V1 will be returned. For a variable V1 with levels "control" and "treatment", a column V1treatment will be returned. In the case of a variable V1 with levels "a", "b", and "c", then the design will contain V1b and V1c columns, each representing the change in ratios relative to reference level "a". For numeric variables, the name will remain the same.
Finally, differential events can be identified by filtering for padj < 0.05, for example.
Lower-level function called within DiffRAC. Loads the libraries required by DiffRAC and verifies the compatibility of the design and count tables. Creates the model matrix and prepares the count tables that the main DiffRAC function will use as input.
DiffRAC.initialize(formula, design, counts_num, counts_denom, mode, bias=1)Please refer to the DiffRAC main function.
Lower-level function called within DiffRAC. Modifies the bias term for a design matrix that is already constructed.
DiffRAC.modifyBias(design_mat, mode, ratio)Please refer to the main DiffRAC function for the other arguments.
The ratio of the new bias constant to the previous bias constant