Bioinformatics

Week 3: Gene Annotation

By Prof. Jackson

NCBI ORF Finder

You can also use the accession ID for search. Since you’re using unknown sequence, you’re trying to annotate, this will not be a good option

Copy and paste your sequence in this window.

Go to: https://www.ncbi.nlm.nih.g ov/orffinder/

ORF Finder parameters

If working with the sequence (genome) that still has mitochondrial you may want to clean your genome and remove the mitochondrial genome. (not applicable for this assignment)

Depending on the organism (if you known it) you may increase or decrease the length. This implies that if the algorithm finds a stop codon before the selected length, it will be ignored until another stop codon that meets the length restriction. For small organism like viruses, bacteria, it may be a good idea to reduce the length from 75.

Some organisms like E.coli have other known start codons than ATG. You may want to identify a cds starting with other codons as well.

ORF finder results

• In most cases the longest of the predicted ORFs is usually the correct ORF but there are times when this may not be true.

• + is for positive strand or forward strand. There are usually 3 frames forward (+) and three frame in reverse (negative (-)) making a total of 6 frames

• The algorithm reads in a sliding window. Frame 1 will start with nucleotide in the first position, frame 2 will start on second position and frame 3 third position. You easily implement your simple ORF predictor with python, Biopython or any other programing language.

Predicted ORFAmino acid sequence matching the ORF

You can click on the Mark flag to select multiple ORFs

Highlight the Orf you would like to include on your selection

Have your mouse on the

graphic for more information

ORF finder results cont.

• BLAST is a validation step that can also be used to further identify your sequence/ gene.

• Ideally you should choose the same database you used when BLASTing to identify your sequence.

• You expect the correct ORF to have similar results as your sequence BLAST results.

• The correct Orf should also have better statistics/scores, i.e., better e-value than the rest.

Select a BLAST database to search BLAST your marked ORFs

BLAST The number of sequences should match the number of ORFs you marked/selected . You can also select the BLAST search database here and change any other parameters you want

ORF BLAST Results

• The default results will be on the first Orf you selected.

• ORF1 seems to have the most hits and acceptable scores, likely to be the correct ORF.

• Selected ORF3 has no results

Use drop down to select results for different ORFs

OFR Finder

• Now that you know ORF1 is the most likely ORF use SmartBLAST instead.

• You can also try searching the multiple ORFs in different databases for further evaluations

Use SmartBLAST

SmartBLAST results • Pay attention to what is suggested as the top hits. They

do not necessarily have all the best scores, as we will examine in the next slides

• Smart BLAST also generates phylogenetic tree.

• In addition, you can perform a multiple sequence alignment

Analyzing selected Best hits results

• They all have same e-value so that can’t be the deciding attribute

• The Homo sapiens has good e-value, no gaps although the identities of 94%

suggest may not be the same organism as the unknown sequence.

• Mus musculus has 4% gaps which is too high compared to others and the

identities is too low to be acceptable as the same organism.

• All the rest have gaps and compatible identities making the homo sapiens the

best hit

Comparing BLAST results

• The first hit seems to have perfect scores and would seem like an obvious choice

• Most of the hits that follow are only predicted, therefore nothing to base on to dismiss the first hit.

• The 14th and 15th hits under homo sapiens quite competing scores and they indicate to be complete. Those should be included for further evaluation.

Click GenBank to go to nucleotide database

Click on each selection for detailed information

Click on gene to link to the gene database for additional information

Further analysis of the BLAST Results

• The top hit indicates that it is Provisional, and the Homo sapiens hit are reviewed.

• For annotation purposes, always choose the reviewed hit.

• This explains why SmartBLAST selected Homo sapiens for the top best hit even though Macaca Mulatta had better scores.

• By now you should be familiar with concepts of Reviewed, Provisional, Model, predicted, etc., for better decisions in selecting hits.

• As covered in week 2, NCBI gene database has graphics that can be used to capture various genomic features.

• Hovering your mouse in the graphics as shown in the figures can get the intron and exon positions.

• You can also download the information by clicking at the end of the download arrows.

On slide 12, you can choose the Ensembl id under gene summary to bring you directly to this page. You can also go to Ensembl website and paste the ID or search with NCBI accession ID

Gene position here. Forward strand is your ORF. Can also be represented as positive or + strand

Select regulation for regulatory regions such as enhancers and promoters

Select transcript to get sequence regions like exons, and introns positions

Select exons to get the coding positions table

Results

Select cDNA to retrieve other sequence region including UTR regions Select variant table for the results below

Results from selecting regulation on the gene tab

Follow the color flag and hover your mouse to see the feature

Results of the regulations features

Same information as Ensembl. UCSC also has all the ENCODE data