HIValign

Purpose: HIValign takes one or more nucleotide sequences in any standard format and aligns them to our curated alignments via the appropriate Hidden Markov Model, or aligns them using multiple sequence alignment by the multiple alignment program MAFFT.

Usage: Paste your sequence(s) into the space provided, or browse for your file. Then, you may choose options and select an appropriate hidden Markov model.

Explanation: below

Input

Paste your sequence(s) [Sample Input]
Or upload your sequence file
Select the HMM Model	HIV-1/SIVcpz HIV-2/Other SIV Let program decide
Select region	Complete Gag Pol Vif Vpr Tat Rev Vpx Vpu Env Nef

Options

Always e-mail result
Align the sequences	By Hmm-align By MAFFT
Codon-Align the sequences	Allow codons to compensate frameshift
Insert a reference sequence	(HXB2 for HIV-1, or SIVMM239 for HIV-2 and SIV)
Translate to amino acid

Note: This program is computationally intensive and may take a while to run. Please don't resubmit your request!

Explanation: HIValign was developed mainly from a implementation of HMMER package. There are five options:

• Check the first option to force your job into background mode; you will receive an email with a result link when the program has finished.
• Check the second option to have the program align your input sequences, in one of two ways:
  • hmm-align aligns your sequences based on our curated DNA alignments HMM model.
  • MAFFT is a multiple-sequence alignment program which can align large sequences. MAFFT is fast, but rough.
If you have not checked the 'align' option, the input sequences must be pre-aligned.
• Check the third option to have the program codon align your sequences based on the reference sequence. A compensating mutation is within 5 codons (default) of an initial frameshift so that the shifted reading frame is left intact.
• Check the forth option to have the reference sequence inserted at the top of your input sequences, before they are aligned.
• Check the final option to have your alignment translated to protein sequence.

HIValign attempts to automatically recognize the format of your input file by using Format Conversion. The result page summarizes the running parameters and information about input, and gives you the alignment of the region which the input touched or you selected , and allows you to download the alignment in different formats, and uses result alignment to do TreeMaker or FindModel program.

Aligning large size sequences is a computationally intensive procedure. Currently, input file which is pre-aligned and smaller than 500 Kb for the user selected model is run immediately; if your input file exceeds the limit, or checks 'Align' option , or uses the model by program decide , your job will be run in batch, and you will receive an email when it has finished. The email contains link to your results.

References:

• HMMER website
• MAFFT (Multiple sequence alignment)
• Eddy SR. Profile hidden Markov models. Bioinformatics 1998.14(9):755-63.
• Gaschen B, Kuiken C, Korber B, Foley B. Retrieval and on-the-fly alignment of sequence fragments from the HIV database. Bioinformatics 2001.17(5):415-8