Alignment methods (MUSCLE, MAFFT)

At the end of today’s session, you

Pre-class work

Read the paper corresponding to your canvas group: MUSCLE, MAFFT

Stop and check

First, let’s take a look at data from last year’s projects here.

Note that you can easily install the two software with BioConda again:

conda install -c bioconda muscle
conda install -c bioconda mafft

In-class activity: MUSCLE and MAFFT

Google slides:

Instructions: Work individually or in teams (people that read the same paper as you). In the slides:

Write down one claim that the authors make about their method (e.g. accuracy, speed, robustness, scalability)
Run the software on the primatesAA.fasta data and write down the command you ran in the slides
- MUSCLE docs
- MAFFT docs
Write a paragraph suitable for a paper Methods section that: a. Describes what you did b. Expliclty connects your choices to the claims described in the paper c. Makes clear what assumptions you are relying on d. Note that you may not say ‘default parameters’ without explaining what that implies in terms of the method

One person per method will present their work to the class.

Let’s open all three in AlignmentViewer:

primatesAA-aligned.fasta        ## clustalw
primatesAA-aligned-muscle.fasta ## muscle
primatesAA-aligned-mafft.fasta  ## mafft

Looking at all three alignments, think of:

Homework: Check out the alignment HW here.

No perfect method
No automatic method
- All methods require manual work of comparing results from different alignment parameters and from different sofware
Take notes of the choices you made and keep track of all comparisons to justify final choice
We probably don’t spend as much time as we should on the alignment step of the phylogenomic pipeline
- We want a blackbox that does not exist yet!

SAGA uses the WSP objective function but uses genetic algorithms instead of dynamic programming (individual=alignment)
very accurate
not scalable

Not a clear answer
Scalability vs accuracy
HW reading: Alignathon
Strategy:
- Run multiple programs and parameter choices
- Filtering is more important than the specific program used (more on filtering later)
- Read program papers and documentation carefully
- Take good note of choices and keep track of all comparisons to justify final choice