-ignore_incomplete_sites flag not 'ignoring' ambiguous sites

[mmelendrez]

Hi- I'am running clonalframeML on a Linux mint distribution 18.2 Sonya

It download and compiled just fine - my alignment has many ambiguous sites, as I was unsure if the algorithm could handle ambiguous sites I ran it without the -ignore_incomplete_sites flag:

mel@lucita ~/Analysis/ClonalFrameML $ ~/Software/ClonalFrameML/src/ClonalFrameML SynA_out_trim_1334_names_MLrooted.nwk SynA_out_trim_1334_names.fasta SynA_out_trim_1334_names_CFML
ClonalFrameML v1.11-3-g4f13f23
Finished reading in control file.

Read 15 sequences of length 1256468 sites from SynA_out_trim_1334_names.fasta
ERROR: FASTA_to_nucleotide(): unsupported base K in sequence 7 (60AY4M2_PEA14) position 397

So that suggested the algorithm doesn't work with alignments that contain ambiguous sites, so I then specified the -ignore_incomplete_sites

mel@lucita ~/Analysis/ClonalFrameML $ ~/Software/ClonalFrameML/src/ClonalFrameML SynA_out_trim_1334_names_MLrooted.nwk SynA_out_trim_1334_names.fasta SynA_out_trim_1334_names_CFML -ignore_incomplete_sites true > test.log.txt

mel@lucita ~/Analysis/ClonalFrameML $ cat test.log.txt 
ClonalFrameML v1.11-3-g4f13f23
ignore_incomplete_sites = true
Finished reading in control file.

Read 15 sequences of length 1256468 sites from SynA_out_trim_1334_names.fasta
ERROR: FASTA_to_nucleotide(): unsupported base K in sequence 7 (60AY4M2_PEA14) position 397

Same error. Suggestions? Does the algorithm only consider 'N' an ambiguous base or does it support the IUPAC ambiguity codes?

Thanks!

[tseemann]

Given you got the error ERROR: FASTA_to_nucleotide(): unsupported base K in sequence 7 it seems very likely that CFML does not support those IUPAC codes.

The manual https://github.com/xavierdidelot/clonalframeml/wiki also says _"The positions listed in the file specified by the -ignore_usersites option lists any site that was not callable or present in all the genomes." so it seems for any NON-CORE sites you need to provide a file of all those coordinates?

[xavierdidelot]

Hi Mel,

Yes indeed ClonalFrameML does not recognise the IUPAC codes, you will need to replace them with gaps. This needs to be done even if the site is excluded using -ignore_user_sites option or -ignore_incomplete_sites because the file is read before these tags are applied.

Best wishes, Xavier

[mmelendrez]

Ok fair enough - perhaps a bit of clarification in the manual on this:

-ignore_incomplete_sites       true or false (default)   Ignore sites with any ambiguous bases.

would be great then. I erroneously assumed this meant if I had ambiguous bases as defined by IUPAC that those sites would get automatically ignored. I chose this over

-ignore_use_sites

option because I didn't have a list of sites to ignore (ie. sites with ambiguities) - as I thought the previous flag would 'auto-ignore' those sites.

I will go back an omit those sites from my alignment and retry. I hesitate to turn them all into 'gaps' as wouldn't that impact genomic reconstruction? Unless gap sites are ignore in the algorithm.

Thanks so much for both speedy replies.

V/R - Mel