The MFold Web Server

1
The MFold Web Server

• By Trevor M. Cickovski
• For CSE598K, Biocomplexity
• Date October 28, 2004

2
Purpose

• To provide applications for secondary structure
  prediction of
• RNA
• DNA
• What is secondary structure?

3
M Multiple Folding

• Original (1980) computed one single min. energy
  folding of RNA
• Multiple Folding algorithm Given RNA
• 1. Predict min. free energy G
• 2. Given a set of possible foldings F1Fn,
  calculate their free energies H1Hn
• 3. Eliminate all foldings i with Hi gt Gg
• g G(P/100)
• P is user defined
• 4. Compute remaining foldings plot each with all
  base pairs.

4
Web Server Details

• Runs on 36 dual-processor Intel/Linux
  workstations
• 1 GHZ processors
• 4 GB RAM
• 73 GB disk
• Can fold 6000 bases in 90 min

5
User Input

• User inputs an RNA or DNA sequence as a character
  string
• All except A-Z, a-z get removed
• Lowercase -gt Uppercase
• For RNA, T-gtU For DNA, U-gtT
• Special Characters
• Folding Constraints
• Unspecified bases B (A), D (C), H (G), and V
  (U/T)
• End-helix pairing W (A), X (C), Y (G) and Z (U/T)
• Wildcards
• R (G/A), Y (C/T), M (A/C), K (G/T), S (G/C), W
  (A/T),
• B (not A), D (not C), H (not G), V (not T), N
  (anything)

6
Input Constraints

• F i j k (Starting at base pair ri-rj, force the
  next k to form)
• P i j k (Starting at base pair ri-rj, prohibit
  the next k from forming)
• F i 0 k (Starting at nucleotide ri, force the
  next k to pair)
• P i 0 k (Starting at nucleotide ri, prohibit the
  next k from pairing)
• P i j k l (Prevent any nucleotide from ri to rj
  from pairing with any from rk to rl

7
Other User Parameters

• Linear or circular sequence
• Folding temperature (Celsius)
• Molar ionic concentrations, DNA only
• P for m-fold algorithm
• Upper bound on fold count
• Window parameter W
• Maximum base pair distance M

8
Parameters for Output

• Image Resolution (60, 72, 110, 200)
• Individual base or outline mode for plots
• Turn on/off grid lines
• Base numbering scheme
• Rotation angle
• Structure annotation (p-num or ss-count)
• Colors and hi-lites

9
Immediate vs. Batch Mode

• lt 800 bases immediate mode
• User waits
• Results removed after 24-26 hrs.
• 801-6000 bases batch mode
• Results sent to the user via e-mail
• Results removed after 48-51 hrs.
• 6001-10400 bases special request

10
The Energy Dot Plot

• Plots
• For energies and structure
• Energy in kcal/mol (1/10)
• Interactable and redrawable
• Show base pair details (i, j, free energy)
• Magnifiable
• In multiple formats (.ps, .png, .jpg, .txt)
• Data Files
• RNAML-formatted secondary structures
• H-num helix values
• Single-stranded counts
• Thermodynamic details

11
Other Servers

• Quikfold
• Zipfold
• Tm
• Hybridization

12
Access

• Open to all users
• Quikfold, zipfold, Tm and Hybridization
• Can be hit by Java applications
• Popular servers

13
Future Work with Server

• Improving accuracy
• Folding/hybridization algorithms
• BLAST algorithm for fast database searching
• Improving user interface
• sir_graph