diff options
Diffstat (limited to 'doc')
| -rw-r--r-- | doc/cli.markdown | 2 | ||||
| -rw-r--r-- | doc/matlab.markdown | 45 |
2 files changed, 41 insertions, 6 deletions
diff --git a/doc/cli.markdown b/doc/cli.markdown index 87df41b..d4a1bb9 100644 --- a/doc/cli.markdown +++ b/doc/cli.markdown @@ -8,7 +8,7 @@ module. ## Quikr\_train ## The quikr\_train is a tool to train a database for use with the quikr tool. -Before running the quikr utility, you need to generate the trained matrix or +Before running the quikr utility, you need to generate the sensing matrix or download a pretrained matrix from our database\_download.html. ### Usage ### diff --git a/doc/matlab.markdown b/doc/matlab.markdown index ca701b9..1e990d9 100644 --- a/doc/matlab.markdown +++ b/doc/matlab.markdown @@ -1,5 +1,4 @@ # Quikr's Matlab Implementation # - The Quikr implementation works in Matlab and also works in Octave, but the Octave version will run much slower @@ -9,16 +8,15 @@ make sure that you are in the quikr's matlab directory (src/matlab/): cd quikr/src/matlab - ### Using Quikr with the default databse ### This is the full path name to your data file: fastafilename='/path/to/quikr-code/testfastafile.fa'; This will give the predicted reconstruction frequencies using the default -training database trainset7\_112011.fa from RDP version 2.4 -Xstar will be on the same basis as trainset7\_112011.fa, so to get the sequences -that are predicted to be present in your sample: +training database trainset7\_112011.fa from RDP version 2.4 Xstar will be on the +same basis as trainset7\_112011.fa, so to get the sequences that are predicted +to be present in your sample: xstar=quikr(fastafilename); @@ -110,3 +108,40 @@ This cell array contains the (unsorted) names of the reconstructed sequences and their concentrations (in the first and second columns respectively) namesandproportions={namescell{:}; proportionscell{:}}; + +### Using Multifasta2otu.m ### + +Usage tips: +* Please name fasta files of sample reads with <sample id>.fa<*> and place them + into one directory without any other f ile in that directory (for example, no + hidden files that the operating system may generate, are allowed in that + direct ory) +* Note: When making your QIIME Metadata file, the sample id's must match the + fasta file prefix names +* Fasta files of reads must have a suffix that starts with .fa (e.g.: .fasta and + .fa are valid while .fna is NOT) +* Modify the top of the Matlab/Octave scripts for <input_directory>, + <output_directory>, <output_filename>, and <train ing_database_filename> + +To use with QIIME, one must run the QIIME conversion tool on our OTU table +output: + + convert_biom.py -i <quikr_otu_table.txt> -o <quikr_otu>.biom + --biom_table_type="otu table" + + +4-step QIIME procedure after using Quikr to obtain 3D PCoA graphs: +(Note: Our code works much better with WEIGHTED Unifrac as opposed to +Unweighted.) + +Pre-requisites: +1. <quikr_otu_table.txt> +2. the tree of the database sequences that were used (e.g. dp7\_mafft.fasttree, + gg\_94\_otus\_4feb2011.tre, etc.) +3. your-defined <qiime_metadata_file.txt> + +The QIIME procedue: + convert_biom.py -i <quikr_otu_table.txt> -o <quikr_otu>.biom --biom_table_type="otu table" + beta_diversity.py -i <quikr_otu>.biom -m weighted_unifrac -o beta_div -t <tree file> (example: rdp7_mafft.fasttree)> + principal_coordinates.py -i beta_div/weighted_unifrac_<quikr_otu>.txt -o <quikr_otu_project_name>_weighted.txt + make_3d_plots.py -i <quikr_otu_project_name>_weighted.txt -o <3d_pcoa_plotdirectory> -m <qiime_metadata_file> |