maxibor / adrsm

Ancient DNA Read Simulator for Metagenomic

Hi there 👋

I am Maxime, a PostDoctoral Researcher in Bioinformatics for metagenomics and ancient DNA, based at the Max Planck Institute for Evolutionary Anthropology in Germany and the Leibniz Institut for Natural Product Research and Infection Biology Hans Knöll Institute.

adrsm's Issues

Support gzipped input/output files

As on the tin :D, in this day and age, uncompressed files are a big no-no 😜

Add option for running offline

Some clusters may have computing nodes that are offline.

@maxibor says that currently adrsm uses:

(Using the JGI API to convert TAXID to species name)

Therefore errors such as

requests.exceptions.ConnectionError: HTTPConnectionPool(host='taxonomy.jgi-psf.org', port=80): Max retries exceeded with url: /tax/pt_name/GCA_013267415 (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x2b5d41b17668>: Failed to establish a new connection: [Errno 110] Connection timed out',))

can happen, if no external internet access is avaliable.

An option to run the tool offline (e.g. using ete3 toolkit) would might be helpful for the above use-case.

Offer option to limit number of output reads

Another factor that could be considered when generating in silico aDNA data is sequencing effort.

Currently adrsm will generate sequencing reads that are equivalent to a 'true sample', i.e., has the original full genomes at the requested genomic depth.

However, sequencing experiments themselves have a fixed limit which is the capacity of the machine (or the amount of DNA actually input into the machine).

I would like to request the option for adrsm to output a user-defined set number of reads. For example, only have 10 million pairs in the output files. To simulate sequencing, these reads should be sampled randomly from within the initially sheared reads (which should have sampling probability represented by varying genome coverage).

Provide informative error when UPAC code hit and tool fails.

Some of the 'genomes' I downloaded and tried to input were actually unfinished assemblies:

e.g. GCA_001057035.1_ASM105703v1_genomic.fna

These contained symbols such as W/R/S in the sequence itself.

I would suggest that either if this is encountered that adrsm reports a nice error (rather than traceback) stating this isn't accepted, and please fix.

You could also suggest the following command to clean it up:

sed -i '/^[^>]/s/[R|Y|W|S|M|K|H|B|V|D]/N/g' *.fna