Skip to main content

Development and Test modes

The following additional options may be available in development versions of LiBiNorm. A ✔ symbol indicates that they are enabled in thecurrent version

Skip Nelder Mead parameter stage ✔

Parameter discovery takes place in two stages within LiBiNorm. The first is to use the Nelder-Mead algorithm to determine an appropriate initial set of parameters for the second stage which uses MCMC to determine the model parameters. The -k option skips the Nelder-Mead stage and chooses random starting points for the MCMC parameter determination.

Set initial values from a file ✔

(Enabled by compling with INITIAL_VALUES defined)

The initial values for the MCMC stage can be preset from values in a file using the -i <filename> option. The format of the file is as follows:

A -0.4246024 0.6783496
B -0.549208 1.12126 -4.833895 -3.512791
C 0.303502 0.2770251 -4.951562
D -0.7436946 1.825406 -3.74649 -3.272332
E -0.4013753 1.874199 -1.150785 -4.402464
BD -0.132114 2.036754 -3.982542 -2.990412 0.8643603

Use genes/transcripts from a list ✔

(Enabled by compling with USE_GENES_FROM_GENELIST defined)

The -g <filename> (S F) option specifies that only the genes or transcripts listed in the sepcified file should be used and all other genes and transcripts described in the gtf or gff file should be ignored. The optional S and F parameters allow just the genes from position S to F to be used.

Specify seed for random number generator (model mode only) ✔

(Enabled by compiling with SELECT_READ_SEED defined)

In each run of LiBiNorm a set of up to 100 forward and reverse reads are selected for each gene. This is done with a random number generator that is initialised with a random seed. The seed can be explicitly specified using the -y N option in order to examine the robustness of the algorithm, by specifically selecting different sets of reads for the parameter determination.

Output debug messages

(Enabled by compiling with OUTPUT_DEBUG_MESSAGES defined)

The -w option causes the program to output additional debug messages.

Pause at end

(Enabled by compiling with PAUSE_AT_END_OPTION defined)

The -x option causes the program to pause at the end waiting for user input rather than simply finishing. This can be useful for debugging.

Output additional genome information ✔

(Enabled by compiling with OUTPUT_FEATURE_DATA defined)

If the -c option is being used, then a further file, <countFilenamePrefix>_genome.txt, is created that provides more information about the information obtained from the feature file. This shows each of the regions that have been identified from the gff or gtf file.

The third and fourth coumn gives the start position of the exon within the RNA transcript and the strandedness of the gene. The final four columns indicate whether the region is part of a gene/transcript being used for parameter estimation, the length of the transcript and the number of forward and reverse reads associated with the whole transcript (not the individual exons).

Extended genome output

Print MCMC run data - model mode only

(Enabled by compiling with PRINT_MCMC_RUN_DATA)

As well as the normal three results data files, this option prints out files for each of the models containing the detailed progress for each of the MCMC runs. It also prints out a file (<resultsFilename>_model_cons.txt) that contains parameter values for the iterations with the lowest (negative) log likelihoods in numerical order.

Additional LiBiNorm modes ✔

(Enabled by compiling with LIBITOOLS defined)

Comparing two landscape files ✔

The LiBiNorm land mode allows the contents of two landscape files to be compared. The command format is:

LiBiNorm land -g <gff_file> -i <attribute> -b <bamfile> <landscapefile1> <landscapefile2>

A single file is generated that combines both landscape files, putting the entries for the same gene one after each other so that they can be compared. The gff file is used to provide coordinates for the genes/transcripts, as identified by attribute so that they can be viewed easily on IGV. The bam file allows the correct chromosome names to be used, the mapping being done based on chromosome length.

Comparing a landscape file and a gene list ✔

The LiBiNorm land2 mode allows the contents of a landscape file to be compared with a gene list. Three output files are produced.

LiBiNorm land2 <landscapefile> <genefile>

  • <filename>.subset.txt is a landscape file that contains the data for the genes/transcripts in the list.
  • <filename>.unused.txt is a landscape file that contains the data for the genes/transcripts not in the list.
  • <filename>.extragenes.txt is a list of the genes associated with the .unused.txt landscape file

Investigate the variation in the log likelihood with a parameter value ✔

The LiBiNorm variation mode calculates the variation in the log likelihood value.

LiBiNorm variation [Options] -N <resultsFile> -p <parameterFile> <landscapeFile>

This calculates the log likelihood for the data in the landscape file (only using the genes that are marked as to be used for parameter finding) and saves the output in <results>_variation.txt.