1 Introduction

The BiocNeighbors package provides several algorithms for approximate neighbor searches:

  • The Annoy (Approximate Nearest Neighbors Oh Yeah) method uses C++ code from the RcppAnnoy package. It works by building a tree where a random hyperplane partitions a group of points into two child groups at each internal node. This is repeated to construct a forest of trees where the number of trees determines the accuracy of the search. Given a query data point, we identify all points in the same leaf node for each tree. We then take the union of leaf node sets across trees and search them exactly for the nearest neighbors.
  • The HNSW (Hierarchical Navigable Small Worlds) method uses C++ code from the RcppHNSW package. It works by building a series of nagivable small world graphs containing links between points across the entire data set. The algorithm walks through the graphs where each step is chosen to move closer to a given query point. Different graphs contain links of different lengths, yielding a hierarchy where earlier steps are large and later steps are small. The accuracy of the search is determined by the connectivity of the graphs and the size of the intermediate list of potential neighbors.

These methods complement the exact algorithms described previously. Again, it is straightforward to switch from one algorithm to another by simply changing the BNPARAM argument in findKNN and queryKNN.

2 Identifying nearest neighbors

We perform the k-nearest neighbors search with the Annoy algorithm by specifying BNPARAM=AnnoyParam().

nobs <- 10000
ndim <- 20
data <- matrix(runif(nobs*ndim), ncol=ndim)

fout <- findKNN(data, k=10, BNPARAM=AnnoyParam())
head(fout$index)
##      [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9] [,10]
## [1,] 9132  486 3702 9226 3669 6981 3108 9122 6643  9331
## [2,] 4647 3732 2324 6478 3408 5033 4518 4443 3730  2800
## [3,] 5338 6931 1751 7620 7500 8486 9228 7420 9835  9790
## [4,] 3731 3099 9586 6848 1331 6740 9061 4712 7945  1727
## [5,] 5027  353  701 5631 7507 9070 9111 8457 3522  2309
## [6,] 8563 4388 9485 5990 9660 1023 6027 2571 5531  6911
head(fout$distance)
##           [,1]      [,2]      [,3]      [,4]      [,5]      [,6]      [,7]
## [1,] 0.7955538 0.7993070 0.8670275 0.8893975 0.8990859 0.9185116 0.9308243
## [2,] 0.8612506 0.8810785 0.8846026 0.9016428 0.9525601 0.9794540 0.9894384
## [3,] 0.8985270 0.9190195 0.9773331 0.9811532 0.9923183 0.9931732 1.0417367
## [4,] 0.8401244 0.8699952 0.8739700 0.8985938 0.8994962 0.9097665 0.9111904
## [5,] 1.0512825 1.0699428 1.0717589 1.0932229 1.1015829 1.1048776 1.1068014
## [6,] 0.7201223 0.8758298 0.9073330 0.9589700 0.9773981 0.9850024 0.9959804
##           [,8]      [,9]     [,10]
## [1,] 0.9489902 0.9665209 0.9728610
## [2,] 0.9967210 1.0027425 1.0114421
## [3,] 1.0571125 1.0683599 1.0922165
## [4,] 0.9245242 0.9503780 0.9547779
## [5,] 1.1175274 1.1209606 1.1239760
## [6,] 1.0304445 1.0322986 1.0459485

We can also identify the k-nearest neighbors in one dataset based on query points in another dataset.

nquery <- 1000
ndim <- 20
query <- matrix(runif(nquery*ndim), ncol=ndim)

qout <- queryKNN(data, query, k=5, BNPARAM=AnnoyParam())
head(qout$index)
##      [,1] [,2] [,3] [,4] [,5]
## [1,] 1346 4530 4406 8088 5497
## [2,] 8659 3643 8990 3902 1001
## [3,] 7100 2624 2153 4897 6453
## [4,] 7749 6135 2230  265 8620
## [5,] 2995 5749 6548  921 1780
## [6,] 7579 8470 2683 8108 7873
head(qout$distance)
##           [,1]      [,2]     [,3]     [,4]     [,5]
## [1,] 0.9977517 1.0536718 1.068333 1.073585 1.077637
## [2,] 1.0567617 1.0594093 1.060906 1.062710 1.075028
## [3,] 0.9056990 0.9921265 1.035131 1.040361 1.051483
## [4,] 1.0141206 1.0184829 1.037667 1.063885 1.068576
## [5,] 1.0487195 1.0592717 1.082760 1.119992 1.138012
## [6,] 0.9695914 0.9891419 1.058603 1.082217 1.083325

It is similarly easy to use the HNSW algorithm by setting BNPARAM=HnswParam().

3 Further options

Most of the options described for the exact methods are also applicable here. For example:

  • subset to identify neighbors for a subset of points.
  • get.distance to avoid retrieving distances when unnecessary.
  • BPPARAM to parallelize the calculations across multiple workers.
  • BNINDEX to build the forest once for a given data set and re-use it across calls.

The use of a pre-built BNINDEX is illustrated below:

pre <- buildIndex(data, BNPARAM=AnnoyParam())
out1 <- findKNN(BNINDEX=pre, k=5)
out2 <- queryKNN(BNINDEX=pre, query=query, k=2)

Both Annoy and HNSW perform searches based on the Euclidean distance by default. Searching by Manhattan distance is done by simply setting distance="Manhattan" in AnnoyParam() or HnswParam().

Users are referred to the documentation of each function for specific details on the available arguments.

4 Saving the index files

Both Annoy and HNSW generate indexing structures - a forest of trees and series of graphs, respectively - that are saved to file when calling buildIndex(). By default, this file is located in tempdir()1 On HPC file systems, you can change TEMPDIR to a location that is more amenable to concurrent access. and will be removed when the session finishes.

AnnoyIndex_path(pre)
## [1] "D:\\biocbuild\\bbs-3.14-bioc\\tmpdir\\RtmpKWPENZ\\file3988b7419e.idx"

If the index is to persist across sessions, the path of the index file can be directly specified in buildIndex. This can be used to construct an index object directly using the relevant constructors, e.g., AnnoyIndex(), HnswIndex(). However, it becomes the responsibility of the user to clean up any temporary indexing files after calculations are complete.

5 Session information

sessionInfo()
## R version 4.1.1 (2021-08-10)
## Platform: x86_64-w64-mingw32/x64 (64-bit)
## Running under: Windows Server x64 (build 17763)
## 
## Matrix products: default
## 
## locale:
## [1] LC_COLLATE=C                          
## [2] LC_CTYPE=English_United States.1252   
## [3] LC_MONETARY=English_United States.1252
## [4] LC_NUMERIC=C                          
## [5] LC_TIME=English_United States.1252    
## 
## attached base packages:
## [1] stats     graphics  grDevices utils     datasets  methods   base     
## 
## other attached packages:
## [1] BiocNeighbors_1.12.0 knitr_1.36           BiocStyle_2.22.0    
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_1.0.7          magrittr_2.0.1      BiocGenerics_0.40.0
##  [4] BiocParallel_1.28.0 lattice_0.20-45     R6_2.5.1           
##  [7] rlang_0.4.12        fastmap_1.1.0       stringr_1.4.0      
## [10] tools_4.1.1         parallel_4.1.1      grid_4.1.1         
## [13] xfun_0.27           jquerylib_0.1.4     htmltools_0.5.2    
## [16] yaml_2.2.1          digest_0.6.28       bookdown_0.24      
## [19] Matrix_1.3-4        BiocManager_1.30.16 S4Vectors_0.32.0   
## [22] sass_0.4.0          evaluate_0.14       rmarkdown_2.11     
## [25] stringi_1.7.5       compiler_4.1.1      bslib_0.3.1        
## [28] stats4_4.1.1        jsonlite_1.7.2