Mgcod | Tool for recognizing genetic codes (incl. multiple genetic codes in phage genomes) and genetic-code- | Genomics library

Mgcod is a Python library typically used in Artificial Intelligence, Genomics applications. Mgcod has no bugs, it has no vulnerabilities, it has a Strong Copyleft License and it has low support. However Mgcod build file is not available. You can download it from GitHub.

In the default mode, Mgcod assigns one genetic code to contig included in a multi-FASTA file. However, by setting the --isoforms flag it searches for evidence of multiple genetic codes. For the prediction of multiple genetic, a sliding window approach is applied. First, a genetic code is predicted based on the highest scoring genetic code model for each window. Then, a dynamic programming approach is applied to merge blocks of windows with the same genetic code that are interrupted by only a few windows with a different genetic code. This is done to avoid the segementation of the genome into too many genetic codes. After blocks with distinct genetic codes are identified, it searched for a change in protein-encoding strand (PES) within the two windows between which the genetic code changes. If a change of PES is found then the intergenic region separating the two genes between which the PES switches is annotated as the switch region. If no PES switch is observed, the intergenic region separating the two windows between which the switch of genetic code occurs is annotated as switch region. The reasoning for using PES information to refine the prediction of genetic code switch region is given by the observed PES bias in phages with multiple genetic codes that coincides with the usage of genetic codes.R1 By default, a window size of 5000 bp (-w 5000) and stride of 5000 bp (-st 5000) is used. We do not recommend to use window size shorter than 5000 bp as the accuracy of Mgcod of determining the correct genetic code declines for shorter sequences. Furthermore, we require at least three consecutive windows with the same genetic code to form a new block (-n 3). We observed that if a genome has multiple genetic codes, the predictions of the two genetic codes models are often indistinguishable in one segment of the genome. Therefore, we annotate those predictions as "isoforms". This is done by comparing the predictions of the genetic code models. i) If model A predicts one long gene that overlaps many short predictions by model B and its coding potential is greater than the accumulated coding potential of the short genes, the prediction of model A is retained and the predictions of model B are dropped. ii) If the predictions of both models are on the same PES and share the stop coordinates or the start coordinates while the stop coordinates differ by less than t nucleotides (by default -t 30), the predictions are considered equivalent and are annotated as isoforms. iii) If the predictions of both models are identical, the gene is dual-coded, but only one prediction is included in the final gene set. iv) In all other cases, the predictions of both models are retained and are labeled as unique. Mgcod can handle multi-FASTA files and outputs one prediction per contig. However, it should be noted that if a multi-FASTA file with contigs from different genomes with different genetic codes is provided, it is recommended to split the sequences into individual fasta files, and analyze them separately using the provided multiprocess_mgcod.py script. The multiprocess_mgcod.py script takes as input the path to a text file with which contains the paths to fasta files to be analyzed, with file path per line.