Count transcripts using an Expectation Maximization (EM) algorithm.

em_count(txreads, txlengths, weights, ntx, nr, maxit = 1000L,
  reltol = 0.01, abstol = 0.01)

Arguments

txreads

A minimal matrix where each row corresponds to an alignment. The first column denotes transcript indices in [0, ntx-1] and the second column denotes read indices in [0, nr -1].

txlengths

The sequence lengths for each transcript. Ideally those should be the effective transcript lengths meaning the overall number of possible alignment start positions in a transcript.

weights

Weights for individual alignments.

ntx

The total number of unique transcripts.

nr

The total number of unique reads.

maxit

Maximum number of EM iterations.

reltol

The relative tolerance for convergence.

abstol

The absolute tolerance for convergence.

Value

A list with the following components.

p

The length-normalized and read scaled transcript counts

iterations

The number of used EM iterations

num_ecs

The number of equivalence classes

change

The last osbserved absolute change in transcript counts