Recently, I have sequenced few samples (seawater source) using Illumina MiSeq platform and good quality sequences (70,000-1,20,000) are taxonomically assigned using QUIIME software pipeline. Where, I could observe 200 to 500 OTUs (Operational Taxonomic Units) per sample. Existing data on similar kind of work have reported more than 1000, even 8000 OTUs.
Will anyone suggest, what could be reason behind obtained results and how to troubleshoot it.
@Rajeev - there would be few things to consider
1. The increased depth could surely be used to cover the less abundant OTUs. But, looking at your data results and your comparison to other study does not go well together, but could be possible. To really compare the data, you need to use exactly same bioinformatic analysis parameters. As the data filtering could be a major reason for the difference in OTUs. Moreover what primers and sequencing they used ?
2. Why do you compare to the number of OTUs with other study ? Is it same sea ? Same spot or same water-depth? Same sampling period or same season?
3. Comparing the overall data and pattern would be better in stead of comparing number of OTUs. This is what I can say as far as I understood the situation
Thank you for your kind advice Sir; do you mean to increase the depth (more number of reads) ?
Hello,
which variable regions were sequenced? Are the same regions from the reference papers? the lenght of high quality reads are the same from the similar data?
Probably the differences that you observed are not only in the sequencing depth, but in the setting of libraries design.
@Rajeev - there would be few things to consider
1. The increased depth could surely be used to cover the less abundant OTUs. But, looking at your data results and your comparison to other study does not go well together, but could be possible. To really compare the data, you need to use exactly same bioinformatic analysis parameters. As the data filtering could be a major reason for the difference in OTUs. Moreover what primers and sequencing they used ?
2. Why do you compare to the number of OTUs with other study ? Is it same sea ? Same spot or same water-depth? Same sampling period or same season?
3. Comparing the overall data and pattern would be better in stead of comparing number of OTUs. This is what I can say as far as I understood the situation
Dear Salvatore La China , I have targeted V3-V4 region and length of paired read was 300 bp. I just mean to say that the number of OTUs, I have obtained are acceptable from paper point of view.
Dear Abhijeet Singh, thank. Moreover, I am using same parameters and conditions for all samples. I just want to know the number of OTUs (200-500/sample), I have obtained are enough or some optimization needed ?
@Rajeev - you are not getting the point... if the samples are originated from different places, its obvious that there are differences in number of OTUs, provided the analysis parameters are same (thats what you claim). And I hope this is what @Salvatore La China is also trying to say.
Dear @Rajeev
I completely avere with @Abhijeet. To compare your data with other you take in cosideration different things.
1 The tools used.
2 The db used.
3 you should use the same normalization method.
4 if you rarefied the data, thake in account that this is not a mandatory step. If you have a consistent variation in terms of number of observation among your samples, consider do not rarefied because in this way you loss a large amount of informations.
5 take in to account also all factors that can influence the sampling.
Dear Rajeev,
I too agreed with the previous answers. In addition, depending upon the method you used for OTU picking can vary the output to a greater extent. Also, when you compare with others, are the samples exactly same? Depending upon the diversity of species in samples, OTUs can vary, and until and unless you are capturing most of the species that are supposed to be present in your samples, those OTUs number might be OK; since you have good number of sequence reads per samples.
Dear
Bishnu Adhikari, Abhijeet Singh and Salvatore La China. Thanks a lot for your valuable information and I really admire your expertise in this field.
Align your unfiltered reads back to the OTUs and check how well they describe the data. I've done a bit of work with soil samples, which is of comparable microbial complexity to seawater, and we typically get around 70 - 95% alignment rate (at 97% similarity) for bacterial OTUs. I haven't used qiime for a few years, but I assume it can output alignment rate when creating an OTU table.
Other than that, as above - there can be many reasons why you get different results to other researchers.
Forgot to mention - rarefaction curves should also give you a good indication of whether sequencing depth is adequate.
- Badges
- Science method