Our Software

  • Active Software. This is software that we actively develop and use. We will be happy to hear about bugs, improvements, etc.
  • Scripts. Useful scripts.
  • Shelved Software. This is software that we no longer support, but we still try to provide working downloads.
  • Proof of Concept. This is experimental code that we created as a proof-of-concept demonstration for different papers. We provide it in case somebody finds it useful, but we do not promise any support if you want to use it.

Active Software


DeepNano is a base caller for MinION, the smallest DNA sequencer currently available. DeepNano is based on recurrent neural networks, is faster, and makes better base calls than manufacturer's Metrichor base caller.

Go to DeepNano page


RNArobo is an RNA structural motif search tool. RNArobo can search sequence databases in FASTA format for a motif defined by a descriptor, which can specify primary and secondary structure constraints. The format of an RNArobo descriptor is an extension of the descriptor format used by RNABob (a search tool by S. Eddy), so that RNABob descriptors are compatible with RNArobo. Unlike RNABob, RNArobo enables you to allow nucleotide insertions in a structural element.

Go to RNArobo page


GAML (Genome Assembly by Maximum Likelihood) is a prototype genome assembly tool based on maximizing likelihood of the assembly in a model encompaasing error rate, insert length and other features of indvidual sequencing technologies. It can combine datasets produced by different technologies (currently Illumina, 454 and Pacific Biosciences).

Go to GAML page


CovEst estimates coverage and size of the genome from NGS data without assembly. It uses k-mer based statistics and works even with datasets with <1x coverage.

Go to CovEst page


PIVO (Phylogeny by IteratiVe Optimization) is a program that reconstructs ancestral gene orders on complex phylogenies. Because it uses the DCJ model, it can handle linear, circular, and multilinear genomes. PIVO is mainly targeted at reconstructing rerrangement histories of mitochondrial genomes.

Go to PIVO page


ExonHunter is a comprehensive gene finder based on hidden Markov models allowing use of variety of additional sources of information (ESTs, proteins, genome-genome comparison, repeats, and more).

Go to ExonHunter page



Script that downloads records from PubMed database and saves them in BibTeX format to use as references in TeX documents.

Go to med2bib page

Shelved Software


TreeArrange is a software that rearranges order of genes in expression arrays to increase visibility of groups of similarly expressed genes.

Go to TreeArrange page

Proof of Concept


Atomizer is a software for automated segmentation of DNA sequences with complex evolutionary history. The goal of this software is to automatically select suitable markers in genomic sequences for reconstruction of evolutionary histories involving large-scale events such as rearrangements or duplications.

Go to Atomizer page


DescAl is a prototype implementation of the descriptor alignment framework developed for sequence-structure search of OB-fold domains in protein databases.

Go to DescAl page


HERD (Highest Expected Reward Decoding) is a new decoding method for HMMs, developed for applications, where is difficult to predict annotation boundaries. Unlike traditional decoding methods (like Viterbi algorithm), HERD scores breakpoints individually to maximize the number of correctly predicted annotation boundaries. Currently our implementation work only on the problem of viral recombination detection in HIV genome.

Go to HERD page


MtConrad is a tool for annotation of genes in mitochondrial genomes, including annotation of group II introns. It is based on conditional random field implementation from popular gene finding package Conrad.

Go to mtConrad page