10 proteins have been predicted as probable drug targets in mycobacterium tuberculosis, while out of 10, only 2 proteins show 20% to 30% similarity
with homosapiens. Can we put it into a drug target group?
Can anybody suggest what should be the essential similarity cut off of pathogen drug target proteins with humans?
@HARESH : actually i have done blast analysis for predicting % identity of target proteins with homosapiens ...it is showing about 20- 30 % identity
Further i am planning to do strucural characterization of target proteins ...as there is no structural data available for it
As per my view you % identiy is not good choice to screen the drug targets, my suggestion is go for e-value of blastp results of each sequence.
The e-value threshold cutoff was set to 0.005 and see this article this may be use full for you.
http://www.ncbi.nlm.nih.gov/pubmed/16213791
http://www.ncbi.nlm.nih.gov/pubmed/9254694
I agree with Babajan B. There can be many factors contributing other than %identity. Check the evalue as mentioned above. if the protein function is known, you can tally it with Humans whether they are matching in case of humans. are there any conserved domains in your proteins? if yes, match with humans. if you are planning to go for a drug designing based on competitive inhibition, better tally the substrates too because substrate analogs may interfere with human systems if pathways are common.
Priyanka, if i get you right, you are targeting proteins in MtB as drug targets that is not similar to Homo sapiens. If this is the case, then the % similarity and the % identity should be as less as possible, because that proteins has to be unique for the organism and not for the humans. So the less the identitiy and similarity the better will be the drug target . I think applying subtractive genomics approach will yield search result.
Best of Luck :)
As Babajan B said, Sequence identity is not the right factor to identify drug targets, which is the concept of Subtractive Genomics approach.
For instance, Two proteins may share less than 30% sequence identity but it may have structural similarity where the drug bind, that makes the drug non-specific which lead to adverse effects.
I suggest you to refer the article of 'TargetDB' by Dr. Nagasuma Chandra in which several criteria has been considered to identify drug targets for MTb.
@ i totally disagree with sarvanan, E- value not only used in Genomic approaches, it used in all sort alingment procedures. and E-value represented as expected probability of alingment occurring by chance. It is a statistical calculation based on the quality of alignment (the score) and the size of the database. For example if an alignment obtained from one database has an evalue of x, the exact same alignment obtained from a database of different size will have an evalue of y.
An evalue of 1e-3 is saying that there is a 0.001 chance that that alignment would exist in the database by chance, that is, if the database contains 10000 sequences, then you might expect that alignment to occur maybe 10 times.
An evalue of 0 is actually a rounded down probability (maybe 1e-250 or something), and is simply saying that there is (almost) no chance that alignment can occur by chance.
The score is the measure of similarity between two sequences, and is calculated from the alignment mat
@ CM Anuradha, I hope you misunderstood the question and my answer for that. I havent discuss about the implication of E-value in biology. I answered that Sequence identity is not the right factor to identify drug targets for MTb, because the question is more specific to MTb. We try to identify drug targets so that i lead to development of new anti-TB drugs real time.
I have used 'Subtractive Genomics approach' and identified 59 drug targets based on sequence analysis (E-value cut-off against human proteome is e10-3) [in press]. Many of the identified drug targets are not satisfactory in the context of 'Cavity analysis', ie, the protein may have similar cavities in human proteome. This rises the possibility of identified potential inhibitors against a specified drug target may have in-intended cross-interaction with proteins from human proteome which may lead to adverse effects.
To quote this again from TargetDB paper, 3611 proteins from MTb dont have close homologues with human proteome in which 59 dont have significant similarity itself. So there is a possibility of 3611 proteins from MTb as drug targets, which is the not the real case. TargetDb suggests several criteria 1. Systems Analysis, 2. Sequence Analysis, 3. Structural assessment, 4. Expression data, 5. Non-reactivity with Anti-targets and 6. Non-reactivity with gut flora. Apart from the above 6 criteria, they also emphasize on 1. expression during persistance, 2. Targets unique to mycobacteria and 3. broad spectrum targets.
Most of the 'in silico' work fail in real time because of improper validation and confirmation bias. To improve the quality of 'in silico' work, One has to really validate them intensively considering the real time scenario.
- Badges
- Science topic
- Similar topics
- Bioinformatics
- Bioinformatics and Computational Biology