HackMD
Mammalia: BDNF notes
Thoughts
Use: https://www.ncbi.nlm.nih.gov/kis/ortholog/627/?scope=7776 for data
~163 species
Use 1 gene per species. Subtranscripts would just be copies of the "main" transcript. May also bias the dataset.
Don't bealign. We can map back to human sites (meaning sites corresponding to the human bdnf) since that is what we are interested in
Do the standard hyphy analyses
Do some lineage searching, break down mammals.
Do some lineage tests.
Check for recombination? with RDP (manual), or GARD (automated) (add to pipeline)?
History of the project
- Google doc of manuscript https://docs.google.com/document/d/1JRxYfG4xq1rP0asyE_h1qzE7FTC8PsqDaOMVJVy-t8U/edit?usp=sharing
- Hackmd, the first one https://hackmd.io/fdPQdyl7RE29VzJJaOmP_g
- Hackmd, for mammalia: https://hackmd.io/DZuuFqk7TaibXig8xYsOLw
Flow of the narrative
Intro, motivation and why this paper is neccessary
results: BUSTEDS for whole gene, MEME for positive sites, FEL for negative sites. No FUBAR, no FADE. aBSREL for branches, if we find interesting branches do lineage report. Decide on and assign lineages/labels. Investigate with CFEL and RELAX.
Done.
Recombination detection: RDP, run the human branches.
FADE: Don't have anything to root the tree on., also its in beta.
FUBAR: results should match FEL, can add to raise confidence on sites.
Can we relate sites with disease, referneces for this.
How to present things.
Tables are mostly go in the supplement (large tables.)
Short tables to focus on key areas within BDNF, highlighting results.
Phylogenetree with clades labelled? Need or not?
Alignment quality, heatmap, not necessary because all of this is heuristic.
^ Filler?
Recombination
Manually tested via RDPv5.5 [ref] with modified settings as follows:
Recombination events where ‘accepted’ in cases more than 2 methods agreed.
Use all models except for LARD.
Slightly modified default parameters…
Sequences are linear
List events detected by >2 methods
Recheck the events
Manually accept events detected by >2 methods.
Alignment was save as a distributed alignment (with recombinant regions separated).
This single fasta was separated into two files the first a recombination free file with 319 sequences (with recombinanted regions separated out) and a second recombinbation files with 21 sequences with these removed sequences from Event 1.
NEED
Figure 1. (MEME, adaptive selection)
Figure 2. (FEL, negative selection)
Figure 3. aBSREL.
Figure 4. Lineages, CFEL/RELAX
etc.
Goal is bioarvix? Strike names.
AGL, MJN
//// Google doc?
