First, I suggest you double check the quality of your 16S rRNA sequence. 86% identity highly implied your strain represented a putative new genus/family if the sequencing was all right. Then you'd better run a phylogenetic analysis to find a placement for your strain. You can also try the Classification implemented in RDP database.

First, I suggest you double check the quality of your 16S rRNA sequence. 86% identity highly implied your strain represented a putative new genus/family if the sequencing was all right. Then you'd better run a phylogenetic analysis to find a placement for your strain. You can also try the Classification implemented in RDP database.

I am no taxonomist, but I guess there is no clear answer to that. Given that metabolic capabilities are easier to aquire and lose than a certain 16S pattern, I would stick with the 16S results.

But: Given that your 16S sequences are high quality, 87% identity indicates that you might have found a new species within the genus, as Fang wrote. That would explain your problems as you cannot identify something with these tests that has not been described before.

BLAST is one of the most updated databases... but 86% similarity clearly indicate that the obtained sequence is probably a new genus...

A 16S rRNA identity as low as 86-87% means that either you have discovered a new bacterial genus or your DNA sequence is full of errors. My personal experience is that the second possibility is more likely, as Fang suggests. Check the quality scores and the chromatograms for your sequence, especially at the positions that differ from other bacteria. Biolog is only useful for confirming that an isolate belongs to a species that is already known, since it relies on a database of typical substrate utilisation patterns for each species. If it is giving an identification with a high level of confidence, it is probably right, but a low-confidence match gives no useful information.

I recommend reading of Stackebrandt.1994 / 10.1099/00207713-44-4-846 . Check the 16S rDNA sequences for chimeras (quality in general) and use AR / SILVA to be sure. Is your strain from the environment or clinical? Cheers,

Martin

It can also depend on the organism, 16S can't differentiate to species levels for some genus such as Bacillus or differentiate between E. coli and Shigella. The database you have for biolog can also play a role, databases are usually first built on reference strains which can differ from Clinical strains and it is worth building up a database with your strains to improve the quality of identifications (though you need to ensure you populate it with validated identifications).

look at the sequencing chromatogram. If your sequencing quality is poor, you will see multiple peaks at a position.

Biolog is a common tool. However, the accuracy is depends on how you grow the cells. Even simple difference is incubation temperature will affect the result.

You may experience some problem if your bacteria is thermophiles or halophile. Please also bear in mind that not all species are covered in Biolog database.

As other I suggest to review the chromatogram and also to consider the sequencing method you use, there are some difference in some base adds or change depending of the sequencing method creating like SNPs so be careful with themes.

Another possibility us the purity of the strain and for that you should start to culture from a single colony.

Add to Young, Biolog or other kinds of test strips could be usefull to identify only narrow range of bacteria. Usually used in identification of pathogenic bacteria. If your strain is known pathogenic one, its relaible, if natural isolate, hard to believe the id result. However, similar result was obtained from rRNA sequence (even homology was very low), your strain may well known one and I strongly recomend check contamination or quality of sequencing rxn.

DNA baser sequence assembler tool may be used to evaluate the sequencer chromatograms. it shows bases health and indicate as trusted and untrusted bases. if all of the bases or high percentage of shows promising result then it should be a new genus provided the fact that Biolog provides the biochemical identification while 16S sequencing is more authentic genetic tool.

Well, I am not a specialist but we have that same problem all the time: We are working mainly with fungi, and when we sequence three genes (18S, ITS and elongation factor alpha) it is almost FRIGHTENING how different the "identifications" are that you get when using each of thes genes independentlyy - can easily be that you get different genera! I think we have to accept that one single sequnece (or even three sequences) is simply not enough for a reliable identification. In the end, you simply obtain the information which is the closest sequence currently present in the database. If somebody submits a new sequence today your "identifaction" (i.e. BLAST hit) can be different tomorrow. By the way, many names that you see in the databases also are not correct. We, for the moment, make the folllowing: sequence 3 or better 5 genes, make a phylogenetic analysis for each of them (if the best hit is species A but the sequence per se in the phylogeny apperas in the middle of a cluster formed by species B we take B) and try to combine sequnece information with classical traits (morphology for fungi., fatty acid patterns for bacteria). I hope these comments are useful, all th best,

Martin

As many people mentioned you should first check the sequence quality before doing BLAST search. As Martin mentioned you can't confirm the identity with single gene sequence. Because it is not always useful for some species and you need to sequence multiple genes to differenciate closely related species. Also you should be careful about the matched sequence in BLAST results. Because you can't trust 100% on all the sequences available in GenBank.

It is much more important the 16S rRNA gene sequence than the BIOLOG result. The BIOLOG is just a multi carbon sources oxidation test. This test needs to be complementary with 16S rRNA gene sequence and other polyphasic aproachs to conclude which species/genus is your strain. If you have the full sequence of 16S rRNA gene you should run a chimera check by Mallard, RDP, arb,... If everything is ok with your sequence, then you need to do a phylogenetic tree to understand what is the closest one. You can see here much more detailed information doi: 10.1099/ijs.0.053454-0 and 10.1099/ijs.0.016949-0.

I hope that my comments are useful for you.

All the best,

Diogo

10.1099/ijs.0.016949-0 is a good reference. One author I know personally, indeed. Dear Mushafau Adebayo Oke, please keep us updated on the outcome. Science is a permanent endeavour! ;-)

I agree with the popular answer. The phylogenetic analysis will help you to make sure the classification of your strain.

I also agree with Diego as it is difficult to correlate these two tests. Neeru Narula

Thanks everyone for your very helpful responses.

Actually, I am not very familiar with molecular methods and bioinformatics. I only ventured here in order to identify my isolates but your responses have spurred my interest further.

On later consultation with some colleagues and after doing some reading, I realized that the sequences are not okay. There are multiple overlapping peaks and the chromatogram looks so unclean compared to the standard ones I have viewed. What bothers me still is that all the isolates (mine and colleagues') had similar pattern of result. We used a universal 16S primer and got clear bands after PCR. I intend to repeat the DNA extraction and probably use another PCR protocol. I only chose the PCR protocol from a paper whose study is related to mine. I am not sure whether that might be appropriate.

Thanks again.

It is quite common to have low quality sequencing for new comers. I assume you sequence the PCR product.

You may want to consider do the PCR with any Taq polymerase and then clone your PCR products into TA-cloning vector, there are a few commercial available brands,

Transform the ligated products into completent E.coli, do a blue white screening, then send for sequencing. The route is longer, but you will get better result. Do remember to keep you E.coli glycerol stock in case you need to redo the sequencing.

BIOLOG is not an proper tool for identification of an organisms. YOu can use this results for substrate utilization of different sugars and others.. you must look 16S rRNA sequence ( minimum of 1400bp ), if the BLAST results showed lower homology like less than 95% propably your strain could be a novel genus. but need to analysis other housekeeping genes and its phylogenetic positions also other chemotaxomonical analysis. IJSEM, SAM, Arch microbiology, jrnof extremophiles jornlas having lot of reference paper for doing these.

I agree with most of the aswers, in particular with Diogo Proença. I would only add a suggestion. The similarity obtained from alignment is very low and this suggests that probably something went wrong with the sequencing. I suggest to check the beginning and the last portion of the sequence, if you are not sure about their quality, probably it is better to trim them out and repeat the alignement using only the high quality portion of the sequence.

A band from a 16S rDNA amplification cannot rule out divergent and diverse micro-organisms in your sample. Either follow the TA cloning sequence approach by sequencing of at least 20 clones or check your sample culture for mixed populations. Good Luck and thanks for the reply to us.

Hi Mushafau

Firstly, I'm no BIOLOG expert, but in general terms comparing BIOLOG results with those of molecular sequencing seems rather like comparing apples and pears. What you are essentially doing is comparing activities with genetics and expecting the same result. You must remember that with BIOLOG there are a limited/finite number of possible results, while microbial diversity on the other hand is considerably greater. It is therefore obvious that many organisms could give the same BIOLOG result. In environmental microbiology at least BIOLOG is not a standard method for organism identification.

As for your 16S rRNA gene sequencing, as others have said I hope the quality of your sequence was good. Assuming that you did have a quality sequence to work with, with few (if any) ambiguities, it would appear that in fact you do have a novel organism. If so, congrats! I know you got 94% coverage, but was this will a full-length gene sequence? If not I'd recommend sequencing the full sequence again. Reproducibility is key. Best of luck :)

The results of 16S rRNA are not enough, just a confirmation. You must have matches with BIOlogs or Bergs Manual Tests or enzymes or criteria which are familiar with your lab

Did you check the BLAST with a Type strain ?

The Blast is not enough for identification. Did you run the phylogenetic tree ?

I trust the Blast for genus determination, but for species, I use a phylogenetic tree with validated sequences from type strains.

Which strains do I need to run a phylogenetic tree with? Is it the closest strains in terms of sequence similarity?

You can go to the updated website http://www.bacterio.net/

Then select your genus. You will find all validated species for each genus. For each species, you will find the type strain with accession number. You have to download the 16 S sequence, for exemple one representaive strain for each speices, perform clustalV or clustalW then build a phylogenetic tree (I use LaserGene) but you can use any tool available on the web (NCBI). Include your strain sequence then you can see the closest species. We perform distance matrix too. Strains from the same genus have usually not more than 5% od dissimilarity. Strains from the same species have usually no more than 1% of dissimilarity. However, some species can not fit in these values. If you need help, do not hesitate...