iRep is a method for determining replication rates for bacteria from single time point metagenomics sequencing and draft-quality genomes.
The method is described in:
"Measurement of bacterial replication rates in microbial communities" - Christopher T. Brown, Matthew R. Olm, Brian C. Thomas, Jillian F. Banfield (Nature Biotechnology 2016).
pip install iRep
Note: iRep requires python 3. For full list of dependencies please see requirements.txt.
iRep -> measure replication rates using draft-quality genome sequences
iRep_filter.py -> combine and/or filter iRep output
bPTR -> measure replication rates using complete genome sequences (modified from Korem et al. Science 2015)
gc_skew.py -> calculate GC skew across complete genome sequences
$ iRep -f sample_data/l_gasseri.fna -s sample_data/l_gasseri*sam -o test.iRep
$ bPTR -f sample_data/l_gasseri.fna -s sample_data/l_gasseri*sam -o test.bPTR.tsv -plot test.bPTR.pdf -m coverage
$ gc_skew.py -f sample_data/l_gasseri.fna
Example output is provided in sample_output.
Running iRep and bPTR requires that you have genome sequences for each of the organisms for which you want to measure replication rates. Each program takes the file paths to each of the genomes, in separate FASTA files, as input. bPTR requires complete (closed) genome sequences, and iRep requires high-quality draft genomes (>=75% complete, <=175 fragments/Mbp sequence, and <=2% contamination). iRep will only make use of scaffolds ≥5 Kbp in length, thus it is important to make sure that estimates of genome completion are calculated based on scaffolds that pass this threshold. Both methods are most accurate when the genomes have been assembled from metagenomes from the samples being studied, or if it is known that an organism is present in the system with a highly similar genome sequence. The second set of inputs are mapping files in SAM format, which should be generated by mapping DNA sequencing reads against assembled genomes using Bowtie2. In this case, provide the path to each SAM file generated for each sample.
Only iRep values passing all genome and mapping quality requirements should be considered as valid. Default thresholds: min cov. = 5, min wins. = 0.98, min r^2 = 0.9, max fragments/Mbp = 175, GC correction min r^2 = 0.0.
Please note that high levels of strain variation will confound results.
iRep and bPTR.py output both a table (.tsv file) of results and a PDF with plots showing genome coverage measurements used for calculating replicaiton rates.
Large SAM files with many unmapped reads will be slower to parse than SAM files that do not included unmapped reads. Unmapped reads can be quickly filtered out of the SAM file using shrinksam.
Users have reported receiving a "ImportError: No module named '_tkinter'" error. This can be fixed by modifying the bash MPLBACKEND variable:
Both iRep and bPTR require ordered SAM files that can be generated using the Bowtie2 --reorder flag. Ordered SAM files, where both reads representing a set of paired reads are ordered one after the other, are requred for these scripts to filter reads based on mapping quality (see -mm option). If the SAM file is not sorted, be sure to either re-run Bowtie with the --reorder flag, or use the --sort option to sort the SAM file; otherwise, the scripts will make incorrect choices about which mapped reads to include when calculating coverage. The --sort option uses Unix sort to re-order the SAM file. This only applies when using a mapping quality cutoff; however, using a quality cutoff is recommend as it helps to prevent off-target read mapping.
See iRepValues.pdf for more information on interpreting iRep results.