protein feature analysis

Question 1

membrane proteins

Answer 1

• single spanning segments or entire protein domains
• membrane region is in a hydrophobic environment
• side-chains tend to be hydrophobic
• main chains cna’t have NH/CO groups that don’t form hydrogen bonds
• single spanning must be alpha helical
• most are alpha helical

Question 2

transmembrane helices

Answer 2

• hydrophobic section ~35-50A wide
• often positive charges in the section abutting the membrane
• each residue in helix advances structure by 1.8A
• usually 20-30 residues long

Question 3

methods to identify TM regions

Answer 3

• early methods
  
  • search for runs of hydrophobic residues
  • scale of hydrophobicity
    
    • Hopp & Woods
    • Kyte & Doolittle
  • hydrophobic plots
• current methods
  
  • machine learning based

Question 4

hydrophobic plots

Answer 4

• moving window of 11 residues
• count hydrophobicity according to scale
• above a cutoff indicates TM region
• originally few sequences known
  
  • improved by sequence availability
  • build up to create HMM for ML

Question 5

mahcine learning methods of TM identification

Answer 5

• HMMs, neural networks or SVM
• OCTOPUS, PHOBIUS

Question 6

homology searching with TM regions

Answer 6

• many TM regions have similar hydrophobic residues
• easy to find false matches for your protein that seem significant
• conservative cutoffs needed
• can be better if only intracellular/extracellular parts of sequence put in to e.g. PSI-BLAST
  
  • reduce false matches

Question 7

signal peptides

Answer 7

• 15-60 residues at start of protein
• directs to correct cellular location
• generally cleaved
• often hydrophobic region followed by pattern (cleavage site)
• SignalP:
  
  • HMMs and neural networks to rpedict location

Question 8

low complexity regions

Answer 8

• composition biased strongly to a small number of amino acids
• in many proteins
• distorts statistical significance of alignments
• SEG
  
  • replaces low complexity with lower case letters in blast

Question 9

coiled coils

Answer 9

• 2/3 intertwined alpha helices
  
  • often hydrophobic where wthey pack together
• short (20 res) or much longer
• identify with COILS

Question 10

disordered proteins

Answer 10

• small disordered regions that can’t be identified experimentally with px/NMR
  
  • often C/N term, or long loop
• or fleixble proteins that adopts multiple structures
  
  • may be structured upon binding
• often lower fraction of hydrophobic residues than folded protein with hydrophobic core
• use neural networks or SVM
  
  • DISOPRED2
  • IUPRED