In philliplab/MotifBinner: Bin reads by Motif

You should be able to point the program at a file and the defaults should allow you to produce a single output file with the consensus reads for each pid that was detected.

Optionally you should be able to access: 1) A FASTA file with all the sequences with no PID 2) A FASTA file of all sequences with PIDs with the names reporting the PIDs 3) A FASTA file with all the discarded outlier sequences where the sequence is named according to its PID 4) A FASTA file with the aligned bins - just a single file 5) Primary result: Consensus sequences

x) A html/pdf report on -- The number of input sequence -- The number of sequence with PIDs -- The lengths of the sequence with/without PIDs -- The number of sequences with no PIDs -- The lengths of the sequences with no PIDs -- The bin sizes before outlier removal took place -- Stats about the number of sequences that were removed as outliers and the effects of this on the intra-bin distances -- Specifically how many bins went from size x to size y for all x and y -- Stats on the number of gaps introduced in the alignment steps -- Stats on the number of ambiguity characters in the sequences. Potentially add in the probabilities of this arising given the read error rate observed in the bin.

All possible parameters and arguments should be available via these batch functions.

The html/pdf report should have switches allowing one to turn on or off certain portions of the report.

The functions: file_to_consensus() read_file() bin_file() loop over pids process_bin() return()

The data structures: original input file the list of bins bin_stats final_consensus sequences

How to report on this? create a big list save this thing to file have a knitr document(s) that can build from this input data

If you call knit2html, then the current environment in R is exposed to the script that is getting knitted, so this provides an easy mechanism for building the reporting structures.

This is now working via the Rscript build_report.R report_file input_data_file mechanism

Ok I have a lot of puzzle pieces now, but I think I should be a little more careful in the way I go about trying to put them all together. Take a step back, breathe, make a design with pen and paper.

philliplab/MotifBinner documentation built on Sept. 2, 2020, 11:41 a.m.