5

SMALT seems to be one of the most used read mappers for bacterial data, see, e.g., this query. I do not say that it is not a great mapper, but I cannot easily see what are its main strengths compared to mappers such as BWA-MEM, Bowtie2, NovoAlign or GEM. Moreover, it is not even published.

Could you name some of its distinguishing features (e.g., user support by Sanger Pathogens)?

So far I have heard only arguments like "We use SMALT because everyone does it.", but this is not convincing enough for me.

terdon
  • 10,071
  • 5
  • 22
  • 48
Karel Břinda
  • 1,909
  • 9
  • 19
  • Is is really that popular? e.g. compare your query to this one: https://scholar.google.com/scholar?as_ylo=2016&q="Bowtie2"+bacteria I suggest you re-phrase to just ask about the differences between SMALT and the other algorithms – Chris_Rands May 31 '17 at 14:09
  • There are big clusters of researchers who do not use anything else than SMALT. For instance, a lot of people in bacterial population genomics. – Karel Břinda May 31 '17 at 14:17
  • 1
    Yes, or this query. In fact, if we go by the results of the three queries, it looks like SMALT is an order of magnitude less common than either Bowtie or BWA. – terdon May 31 '17 at 14:34
  • @terdon BWA-MEM has 617 results. SMALT 162. I agree that it is less but it is still a highly used mapper and in certain communities the dominant one. Your last query (BWA) has a lot of results like biological warfare agents, etc. – Karel Břinda May 31 '17 at 14:39
  • Sure, counting google results is a horrible way of doing this. My point is that I don't see much evidence that this aligner is all that popular. I hadn't even heard of it, for example, and I've been working on NGS for the past couple of years and around people working on NGS for more than 15. I think you just happen to be in a community where this tool is over represented. I would have told you the de facto standard is BWA instead. I guess it just depends on what one has been exposed to. – terdon May 31 '17 at 14:49
  • I completely agree – I am now in a community where SMALT is highly overrepresented. I also consider BWA-MEM to be a standard (it is a little bit unfair for other mappers to talk about BWA as a whole since it comprises three distinct mappers). This is why I am trying find some reasons why SMALT is so popular in certain communities and to understand its distinguishing features from "more standard" mappers. – Karel Břinda May 31 '17 at 15:00
  • Heh, I refer to BWA because where I work we've found that bwa sampe + bwa samse gives better results than bwa mem :) – terdon May 31 '17 at 15:35
  • @user172818 Thank you!! This is exactly what I wanted to know. – Karel Břinda Jun 01 '17 at 15:16
  • Btw. What is interesting about the up/down-votes here is that whenever this question got 1 like, someone immediately put 1 dislike so the sum was 0 most of the time. – Karel Břinda Jun 01 '17 at 15:17
  • 1
    @user172818 could you please put your flexibility explanation in as an answer to this question; it seems more appropriate as an answer rather than a suggestion for improvements to the question. – gringer Jun 02 '17 at 07:46
  • It is also recommended mapper for reapr, because it is able to align mate pairs independently to each other (-x option). I do not really know if other mappers do provide this option as well, I have not investigate that... – Kamil S Jaron Jun 04 '17 at 18:02

1 Answers1

3

I have used smalt a couple of times. Smalt is more flexible than bwa/bowtie and can be tuned to be more sensitive to divergent hits, which is useful to certain applications. I heard Sanger did evaluate a few mappers for some non-typical applications (not sure what). They found smalt to perform better. Also smalt was developed at Sanger. I guess they could modify smalt based on their needs, though I don't know the details.

user172818
  • 6,515
  • 2
  • 13
  • 29
  • Thank you for your answer. Btw, Do you known how SMALT compares to NovoAlign? My guess is that Novoalign could perform better, but I have no evidence for it. – Karel Břinda Jun 05 '17 at 21:35